US20120294521A1 - Method for 3d video content generation - Google Patents
Method for 3d video content generation Download PDFInfo
- Publication number
- US20120294521A1 US20120294521A1 US13/456,279 US201213456279A US2012294521A1 US 20120294521 A1 US20120294521 A1 US 20120294521A1 US 201213456279 A US201213456279 A US 201213456279A US 2012294521 A1 US2012294521 A1 US 2012294521A1
- Authority
- US
- United States
- Prior art keywords
- value
- region
- color feature
- feature region
- texture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 233
- 230000008569 process Effects 0.000 claims abstract description 136
- 238000003709 image segmentation Methods 0.000 claims abstract description 13
- 239000013598 vector Substances 0.000 claims description 150
- 238000004422 calculation algorithm Methods 0.000 claims description 41
- 238000012360 testing method Methods 0.000 claims description 24
- 230000011218 segmentation Effects 0.000 claims description 14
- 239000003086 colorant Substances 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 4
- 238000009966 trimming Methods 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 14
- 230000004438 eyesight Effects 0.000 description 6
- 210000003128 head Anatomy 0.000 description 6
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000001747 exhibiting effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010998 test method Methods 0.000 description 2
- ZYECOAILUNWEAL-NUDFZHEQSA-N (4z)-4-[[2-methoxy-5-(phenylcarbamoyl)phenyl]hydrazinylidene]-n-(3-nitrophenyl)-3-oxonaphthalene-2-carboxamide Chemical compound COC1=CC=C(C(=O)NC=2C=CC=CC=2)C=C1N\N=C(C1=CC=CC=C1C=1)/C(=O)C=1C(=O)NC1=CC=CC([N+]([O-])=O)=C1 ZYECOAILUNWEAL-NUDFZHEQSA-N 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 210000000887 face Anatomy 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/50—Depth or shape recovery
- G06T7/55—Depth or shape recovery from multiple images
- G06T7/579—Depth or shape recovery from multiple images from motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Definitions
- the present invention relates to a method for generating three-dimensional video content particularly to a method that is capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- the two main and commonly adapted 3D imaging technique may be categorized into two methods: one being a array-based filming technique, the other being a depth-based filming technique.
- array-based filming technique involves setting a number of video cameras in an array arrangement, where its general practice involves using two video cameras.
- these two video cameras should undergo a corrective process before filming starts so as to ensure that there is no horizontal position displacement issue between these two video cameras.
- the distance between these two video cameras should emulate the distance between a normal person's eyes (i.e. the visual parallax of the eyes), so as to simulate a possible perspective of a normal person.
- the requirement for these two video cameras to be maintained in a synchronous state at a run-time setting incurs extraneous effort and time during use of the equipments.
- these two video cameras may be subject to changes in light abundance or changes in internal parameters of the video cameras, making the three-dimensional images so captured open to discoloring.
- depth-based filming technique turns to use one traditional video camera while also coupling with another depth video camera for filming purposes.
- the aforementioned depth video camera operates on a “time of flight” principle, working by means of calculating the time period from the emitted infrared light hitting the desired object to the same infrared light returning back into the depth video camera, so as to determine the actual distance between the desired object for phototaking and the video camera.
- the relative distance between the aforementioned traditional video camera and the aforementioned depth video camera may be subject to adjustment prior to filming begins so as to ensure that the pixels picked up by the traditional video camera are all well aligned with their corresponding depth value.
- the depth resolution offered by the depth video camera still faces serious limitation, and it is not possible to provide a depth image of high image resolution, which altogether creates room for improving the image quality of composite three-dimensional image.
- the underlying issue of increased operating cost with these setups remain to be of concern as, for example, the array-based filming technique requires a multiple number of video cameras to be on standby during filming.
- another disadvantage hereof is that the three-dimensional image produced therefrom will be more open to distortion and become useless if each participating video camera is not adjusted relative to its position from one another or does not synchronize accordingly.
- Another disadvantage associated with the known setups is of the high price for a depth video camera, which has demonstrated to be a barrier for higher market penetration, making it less possible to effectively lower operating cost for producing a three-dimensional video content with a depth video camera.
- two-dimensional image signal is the type of image signal that is most commonly used in the current relevant state of art. It will therefore be understood that the issues encountered by the aforementioned ways for obtaining three-dimensional image content (three-dimensional image signal) will be overcome if a new way for converting two-dimensionality into three-dimensionality is found.
- One object of the present invention is to provide a method for generating three-dimensional video content particularly to a method that is capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- a method for generating a three-dimensional image signal comprising the following steps: (A) receiving a 2D image and generating a region of interest distribution map from the 2D image, wherein the two-dimensional image signal comprises a plurality of individually colored image pixel, and the region of interest distribution map comprises a region of interest; (B) executing a color feature capture process, for forming a plurality of color feature regions; (C) executing an image segmentation process basing on the texture feature of the plurality of color feature regions, for forming an image region distribution map; (D) executing a depth map generation process, for generating a depth map basing on the region of interest distribution map and the image region distribution map, wherein in the depth map each color texture feature region is individually characterized by a depth value; (E) executing a 3D image generation process, for forming the 3D image basing on the image region distribution map and the depth map; and (F) chaining a plurality of the 3D images to form
- FIG. 1 illustrates a flow chart of steps during operation of the three-dimensional image content generation method according to an embodiment of the present invention
- FIG. 2 illustrates a flow chart of steps during operation of the color capture process of the three-dimensional image content generation method according to an embodiment of the present invention
- FIG. 3A is a schematic view illustrating the process of respectively mapping the colors of two image pixels contained in a two-dimensional image signal to a color plane of an HSI color space, for respectively clustering the colors of the two image pixels into a color region or a fuzzy region;
- FIG. 3B is a graph showing the result from the clustring as illustrated in FIG. 3A according to the present invention with respect to the process of designating the first image pixel with a color feature vector;
- FIG. 3C is a graph showing the result from the clustring as illustrated in FIG. 3A according to the present invention with respect to the process of designating the second image pixel with a color feature vector;
- FIG. 4 illustrates a flowchart of steps during operation of an image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention
- FIG. 5 is a diagram illustrating the disposition of twelve two-dimensional Gabor filters included in a Gabor filter package on a spatial plane, operational in the image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention
- FIG. 6A shows a distribution diagram for a plurality of color feature regions represented in a two-dimensional image signal
- FIG. 6B shows a distribution diagram for a plurality of color feature regions of FIG. 6A as further segmentized after a re-segmentation process
- FIG. 7A illustrates a flow chart of steps during operation of an image segment merging process according to an embodiment of the present invention
- FIG. 7B is a distribution diagram for a plurality of texture color feature regions represented in a two-dimensional image signal
- FIG. 7C is an illustrative diagram demonstrating the result of merging a texture color feature regions of tag number 1 and a texture color feature regions of tag number 5 for deriving a transitorily merged texture color feature regions;
- FIG. 8 illustrates a flow chart of steps during operation of a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention
- FIG. 9A illustrates a flow chart of steps during operation of generating a graph showing distribution for region of interest of the three-dimensional image content generation method according to an embodiment of the present invention
- FIG. 9B is an illustrative diagram showing distribution of nine default image pixels of an image signal
- FIG. 9C shows a flow chart of steps during operation of calculating for each of a plurality of image pixels and motion feature values of a plurality of default image pixels
- FIG. 9D is a distribution diagram for five texture color feature regions and one region of interest in a two-dimensional image signal
- FIG. 9E is an illustrative diagram exhibiting the result of designating equal depth values among four texture color feature regions according to the five texture color feature regions of FIG. 9D ;
- FIG. 10A is a diagram showing distribution of a plurality of texture color feature regions of an image region
- FIG. 10B is a flow chart of steps during operation of an aspect ratio value capture procedure required for a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention
- FIG. 10C is a diagram demonstrating a rectangular frame enclosing a portion of the texture color feature region, and cropping the edges of the texture color feature region;
- FIG. 10D is a diagram showing an adjusted rectangular frame enclosing a 70% area of the texture color feature region as presented after the size and position of the rectangular frame onto a two-dimensional image signal is adjusted;
- FIG. 11A is a diagram displaying designation of an identical depth value for a plurality of image pixels
- FIG. 11B is a diagram displaying designation of a linear distribution of depth values for a plurality of image pixels
- FIG. 11C shows a chart exhibiting the linear relationship corresponding to a depth value for use in attributing a pending depth value to an image pixel
- FIG. 12 is another embodiment for a flow chart of steps during operation of depth map generation process for an adjusted depth map operable to produce a corresponding image signal;
- FIG. 13A shows a chart of sub-steps for the step F3 of the depth map generation process according to an embodiment of FIG. 12 ;
- FIG. 13B is a diagram showing the relationship between each of the parameters used in generating an anterior depth map based on a forward motion vector and a backward motion vector;
- FIG. 14 illustrates a flow chart of steps during operation of the three-dimensional image formulation process of the three-dimensional image content generation method according to an embodiment of the present invention.
- FIG. 1 shows a flow chart of steps during operation of the three-dimensional image content generation method according to an embodiment of the present invention.
- An advantage of the three-dimensional image content generation method may be for transforming a two-dimensional image signal into a three-dimensional image signal.
- the method for generating a three-dimensional image content comprises the following steps:
- Step (B) of the three-dimensional image content generation method of the present invention may be referred to FIG. 2 , which is directed to a flow chart of steps during operation of the color capture process of the three-dimensional image content generation method according to an embodiment of the present invention, wherein the steps of which comprise the following:
- one two-dimensional image signal includes a plurality of image pixels, and these image pixels each contains a color.
- the two-dimensional image signal format is unlimited, which may be YcbCr420 or RGB444.
- the image signal is mapped to a color space model, making the colors contained by each image pixels projected individually to a pair of corresponding points in the color space model, and the coordination for the corresponding point is composed of a first parameter value, a second parameter value, and a third parameter value.
- the color space model may be of any desirable type, which may be a HSI color space model, a RGB color space model, a YUV color space model or a CMY color space model.
- FIG. 3A a schematic view is provided illustrating the process of respectively mapping the colors of two image pixels contained in a two-dimensional image signal to a color plane of an HSI color space, for respectively clustering the colors of the two image pixels into a color region or a fuzzy region.
- the aforementioned color feature capture process works to divide a color plane of the HIS color space model into five color regions, each representing a first color region, a second color region, a third color region, a fourth color region and a fifth color region, wherein each color region individually responds to a particular color.
- the first color region corresponds to a red region 31 (red)
- the second color region corresponds to a green region 32 (green)
- the third color segment corresponds to a cyan region 33 (cyan)
- the fourth color region corresponds to a blue region 34 (blue)
- the fifth color segment corresponds to a magenta region 35 (magenta).
- the contained angle in red region 31 is between 350° to 410°
- the contained angle in green region 32 is between 95° to 120°
- the contained angle in cyan region 33 between 170° to 190°
- the contained angle in blue region 34 is between 230° to 250°
- the contained angle in magenta region 35 is between 290° to 310°.
- a fuzzy region is defined as a space located between two neighboring color regions (those not covered by the five color regions as mentioned above). For example, a region located between a first color region and a second color region, or a region located between a third color region and a fourth color region. Therefore, as shown by FIG. 3A , the color plane has five color regions and five fuzzy regions.
- the two image pixels contained by the aforementioned image signal is each a first image pixel and a second image pixel, and these two image pixels each has a color, as previously recited by step (A1), the color feature capture process.
- the aforementioned image signal is mapped to a HSI color space model, making the colors contained by first image pixel and second image pixel mapped to a corresponding point in the HSI color space model.
- the corresponding point P 1 for the color of the first image pixel in the HSI color space model is (105°, 0.75, 90)
- the corresponding point P 2 for the color of the second image pixel in the HSI color space model is (130°, 0.625, 85)
- the first parameter is for a hue value
- the second parameter is a saturation value
- the third parameter is an intensity value.
- step (A3) the next step is to compare the third parameter value of the corresponding point P 1 and P 2 , which is the intensity value, with an adjustment threshold value, and then work out an adjusted second parameter value based on a result of the comparison and their respective second parameter value, which is again, saturation value.
- an adjusted threshold value is between 70 and 90, in an even more preferred embodiment, the value is between 80 and 87.
- the aforementioned adjusted threshold value is 85.
- S′ is an adjusted saturation component
- S is a saturation component
- the abovementioned formula can further modified as:
- S′ is an adjusted saturation value
- S saturation value
- T is the adjusted threshold value
- the above-mentioned formula can be further modified as:
- the adjusted threshold value T is 85, the adjusted threshold value T may also be other values, such as 90 or 80.
- the corresponding point P 1 for the color of the first image pixel in the HSI color space model is (105°, 0.75, 90), and the corresponding point P 2 for the color of the second image pixel in the HSI color space model is (130°, 0.625, 85).
- the intensity value (90) of P 1 is higher than the adjusted threshold value (85)
- the adjusted saturation value will work out to be 0.8 as calculated by the aforementioned formula 2.
- the coordination of P 1 in FIG. 3A on a color plane is (105°, 0.8).
- the adjusted saturation value will work out to be 0.8 as calculated by the aforementioned formula 4.
- the coordination of P 2 in FIG. 2A on a color plane is (130°, 0.7).
- step (A4) of the aforementioned color feature capture process the colors contained by two image pixels are categorized into one of the five color regions or one of the five fuzzy regions of a color plane as shown in FIG. 3A , wherein the two image pixels are determined from the first parameter and the adjusted second parameter of the corresponding point P 1 and P 2 respectively, which in other terms may be referred to as the hue value and the adjusted saturation value.
- the corresponding point P 1 is located at (105°, 0.8), and can be categorized into the green region 32 , and because the corresponding point P 2 is located at (130°, 0.7), it is categorized into the fuzzy region 36 between the green region 32 and cyan region 33 .
- a color feature vector is designated to each of these image pixels according to a result of the abovementioned categorization.
- the color feature vectors designated to the image pixel include a first vector, a second vector, a third vector, a fourth vector, a fifth vector, and a sixth vector.
- the first vector responds to a red region 31
- the second vector responds to a green region 32
- the third vector responds to a cyan region 33
- the fourth vector responds to a blue region 34
- the fifth vector responds to a magenta region 35
- the sixth vector responds to an achromatic region.
- the value of the sixth component is equal to 1 subtracting adjusted saturation value (i.e. 1 ⁇ S′), while also the sum of the values for the first vector, second vector, third vector, fourth vector, fifth vector, and sixth vector is 1.
- the sixth vector are of non-zero values, and the sum of the two values is 1.
- the vector of which the color of the corresponding image pixel is categorized to the color region (i.e. the second component) are of non-zero values, and the sum of the two values is 1.
- the achromatic color feature relates to the amount of white light component within a color of an image pixel.
- the amount of such component can enable a viewer see the difference between a color's saturation level.
- human perception to an image pixel may differ significantly if the colors of which share the same hue but different level of saturation.
- the hue components of bright red or dark red may be very similar, but differ significantly with respect to saturation; this is because the white light components differ by a significant degree, which therefore has stimulated various sources defining achromatic color feature as for aiding in capturing features having improved color properties.
- the color feature vectors designated to the image pixel will include a first vector, a second vector, a third vector, a fourth vector, a fifth vector, and a sixth vector.
- the value of the sixth vector is equal to 1 minus the adjusted saturation component, and its mathematical representation may be 1 ⁇ S′.
- the sum of the values for the first vector, second vector, third vector, fourth vector, fifth vector, and the sixth vector is 1.
- only two vectors that respectively correspond to the color regions on the two sides of the aforementioned fuzzy region i.e. the second vector and the third vector
- the sixth vector are of non-zero values, and the sum of the three values is 1.
- C is a value of the second vector
- S′ is the adjusted saturation component
- Belongingness is a degree of belongingness
- the degree of belongingness may be determined by:
- U is an angle contained by a distant border of a color region on the color plane, to which the second vector corresponds
- L is an angle contained between a neighboring border of a color region on the color plane, to which the second vector corresponds
- H is an angle contained between the hue component and the color plane.
- the fuzzy region's belongingness to the green region 32 is:
- the required parameters for calculating P 2 's belongingness with respect to the cyan region 33 will be: an angle contained by one distant border of the cyan region 33 on the color plane (U) is 120°, and one angle contained by one neighboring border of the cyan region 33 on the color plane (L) is 170°. Accordingly, the belongingness of P 2 with respect to the cyan region 33 may be calculated as follows:
- a clustering algorithm is used to cluster image pixels into clusters based on the designated color feature vector of the image pixel.
- the cluster algorithm herein may be K-means cluster algorithm methodology, CRLA (constraint run length algorithm) cluster algorithm methodology, SBKM (symmetry distance based K-means algorithm) cluster algorithm methodology, etc., but the preferred methodology is K-means cluster algorithm.
- the color feature capture process of step (B) of the three-dimensional image content generation method of the present invention is capable of precisely capturing a color feature vector from the colors of a two-dimensional image signal having a plurality of image pixels, such as V 1 (0,0.8,0,0,0,0.2) and V 2 (0,0.56,0.14,0,0,0.3), and then designating these two color feature vectors to their corresponding image pixels.
- V 1 (0,0.8,0,0,0,0.2)
- V 2 (0,0.56,0.14,0,0,0.3
- the image pixels of a two-dimensional image signal can be categorized into a plurality of color clusters based on the color feature vector designated to each image pixel.
- a plurality of neighboring and being-of-same-color-cluster image pixels are included in the same color feature region, thereby making it possible to form a plurality of color feature regions.
- FIG. 4 illustrates a flowchart of steps during operation of an image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention, wherefore the process includes the following steps:
- the two-dimensional image signal comprises a plurality of image pixels.
- the choice of format for the two-dimensional image signal and such lack of limitation allows a two-dimensional image signal represented by any format to be used in the image segmentation process using texture feature.
- a Gabor filter is installed with the two-dimensional image signal to implement a Gabor filtering process to each image pixel, and subsequently execute a value algorithm process with respect to the output from the execution of the Gabor filtering process.
- value algorithm process is for calculating the modulation magnitude outputted from the Gabor filtering process.
- the Gabor filter package includes an internal subband Gabor filter 51 and an external subband Gabor filter 52 , and the internal subband Gabor filter 51 and external subband Gabor filter 52 each includes six two-dimensional Gabor filters, meaning that each of the internal subband Gabor filter 51 and the external subband Gabor filter 52 has an identical number of two-dimensional Gabor filters.
- FIG. 5 A schematic view of the distribution of the twelve two-dimensional Gabor filters on the space plane is illustrated in FIG. 5 .
- the twelve two-dimensional Gabor filters can each be described by the following formula:
- each ⁇ x , ⁇ y each is a Gaussian kernel standard deviation
- ⁇ is a Fourier fundamental frequency
- ⁇ is a direction of the two-dimensional Gabor filter.
- the standard deviations for the six two-dimensional Gabor filters of the internal subband Gabor filter 51 , ⁇ x , ⁇ y are preferred to be configured at between 1.75 and 6.5, and the standard deviations for the six two-dimensional Gabor filters of the internal subband Gabor filter 51 , ⁇ x , ⁇ y are preferred to be configured at between 1.75 and 4.5.
- the Fourier fundamental frequency ⁇ is preferred to be configured at between 0.2 ⁇ and 0.6 ⁇ .
- the next step (B3) involves individually designating each of the image pixel of the two-dimensional image signal a texture feature vector in accordance with the output (modulation magnitude) obtained from executing the aforementioned value algorithm process.
- the number of dimensions of these texture feature vectors is equal to the number of the two-dimensional Gabor filters used in the Gabor filtering process in step (B2).
- step (B4) a segmentation process is implemented on the two-dimensional image signal according to the texture feature vector designated to each image pixel, such that the two-dimensional image signal has a plurality of texture feature regions. Additionally, the plurality of image pixels in the same texture feature region each has a texture feature vector of identical cluster. Because of this, the two-dimensional image signals that are results of a segmentation process are segmented into a plurality of texture color feature region, and these texture color feature regions can be distributed to any location within the entire area realized contained in the two-dimensional image signal.
- the segmentation process in the present embodiment uses a K-means cluster algorithm. Since K-means cluster algorithm is widely used in numerous fields of technology and is commonly known by people skilled in the relevant art, an elaboration of the K-means cluster algorithm is omitted.
- Step (B5) involves individually executing a second segmentation process in accordance with distribution of the texture feature region in the two-dimensional image signal, such as their relative positions in the two-dimensional image signal, such that at least one color feature region has a plurality of texture color feature region.
- the second segmentation process herein involves the following steps:
- FIG. 6A shows the distribution of a plurality of color feature regions in a two-dimensional image signal
- FIG. 6B shows a resulting situation after a second segmentation process is finished, wherein the color feature region of FIG. 6A is further segmented into a plurality of texture color feature regions.
- step (B6) involves forming an image region distribution map obtained from step (C) of an embodiment of the three-dimensional image content generation method of the present invention, based on the texture color feature region in the two-dimensional image signal.
- step (C) can further include a step of executing an image region combination process subject to actual requirement, such that a plurality of texture color feature region is combined into a combined image texture color feature region.
- FIG. 7A is directed to an image region combination process of an embodiment of a three-dimensional image content generation method, and the image region combination method includes the following steps:
- (C3) according to an order arranged from the texture color feature region having highest area value to the texture color feature region having lowest area value, executing a combination test process on each texture color feature region so as to form a transitory combined image texture color feature region, the transitory combined image texture color feature region includes the texture color feature region and a neighboring texture color feature region, and capturing an area value and a perimeter value of the transitory combined image texture color feature region so as to determine a closeness value of the transitory combined image texture color feature region;
- (C4) comparing the closeness value of the transitory combined image texture color feature region and the closeness value of one of the texture color feature region, and when the closeness value of the transitory combined image texture color feature region is higher than the closeness value of one of the texture color feature region, setting the transitory combined image texture color feature region as a combined image texture color feature region.
- FIG. 7B and FIG. 7C are referred for detailing the aforementioned combination test process.
- FIG. 7B displays the way a plurality of texture color feature region is disposed in a two-dimensional image signal
- FIG. 7C illustrates combination of a texture color feature region labeled as 1 and another texture color feature region labeled as 5 to give rise to a transitory combined image texture color feature region.
- the closeness value of a plurality of texture color feature region can be determined using the following formula:
- C closeness value
- A area value
- P perimeter value
- the aforementioned image region combination method executes a combination test process on each texture color feature region so as to form a transitory combined image texture color feature region, wherein the transitory combined image texture color feature region includes one of the texture color feature regions and another, neighboring texture color feature region. Afterwards, the area value and perimeter value of the transitory combined texture color feature region are determined and ultimately the closeness value is determined accordingly.
- the aforementioned combined test process is executed starting with the texture color feature region labeled as 1.
- such arrangement begins with the plurality of texture color feature regions neighboring the texture color feature region labeled as 1, for example, combining the texture color feature region labeled as 1 and the texture color feature region labeled as 5 into a transitory combined image texture color feature region, then, combining the texture color feature region labeled as 1 and the texture color feature region labeled as 4 into another transitory combined image texture color feature region, and the above process continues on further.
- a threshold value identification process will be executed before the aforementioned “combination test process” is executed, meaning that the area value and closeness value of each of the texture color feature region are compared respectively with an area threshold value and a closeness threshold value.
- the aforementioned combination test process is only executed when the area value of texture color feature region resulting from the comparison is lower than the aforementioned area threshold value, and when the closeness value of the texture color feature region is lower than the closeness value of the aforementioned threshold value.
- the area threshold value is between 120 to 240, wherein the preferred embodiment is 180; the closeness threshold value is between 0.001 to 0.008, wherein the preferred embodiment is 0.003, depending on the environmental conditions.
- the texture color feature region labeled as 1 and the texture color feature region labeled as 5 are combined into a transitory combined image texture color feature region, then the area value and perimeter value of the transitory combined image texture color feature region obtained the combination of the texture color feature regions labeled as 1 and 5 can be determined. Accordingly, the closeness value of the transitory combined image texture color feature region can be determined.
- the closeness value of the transitory combined image texture color feature region and the closeness value of the transitory combined image texture color feature region labeled as 1 are compared.
- the closeness value of the transitory combined image texture color feature region is higher than the closeness value of the texture color feature region labeled as 1
- the transitory combined image texture color feature region is set as the combined image texture color feature region.
- the closeness of the texture color feature region is not higher than the closeness value of the texture color feature region labeled as 1
- the aforementioned process is repeated.
- An instance illustrating this embodiment is to combine the texture color feature region labeled as 1 and the texture color feature region labeled as 4 to form a transitory combined image texture color feature region, and then capture the area value and perimeter value of another transitory combined image texture color feature region to further determine the closeness value of this transitory combined image texture color feature region.
- the previously mentioned ordering method (arranging the texture color feature regions from a texture color feature region having highest area value to a texture color feature region having lowest area value) is replaced with designating the texture color feature region of second largest area value to be the standard texture color feature region for performing the combination test procedure.
- FIG. 8 relates to a flow chart of steps during operation of a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention:
- the aforementioned aspect ratio test process is used for calculating the aspect ratio value of the texture color feature region, and when the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, the depth value designated to the texture color feature region is equal to the depth value designated to another texture color feature region that is perpendicularly neighboring to the former.
- the aforementioned overlapping rate algorithm process for the region of interest is used in calculating a overlapping rate for between a texture color feature region and a region of interest (including the distribution map of region of interest of step (A) of the present invention), and all the texture color feature regions that are neighboring each other and also have overlapping operation that is higher than a overlapping operation threshold value are designated the same depth value.
- FIG. 9A is a flow chart of steps during operation of generating a graph showing distribution for region of interest of the three-dimensional image content generation method according to an embodiment of the present invention:
- the number of standard image pixels of the group for constituting a region of interest is smaller than the number of standard image pixels of the group for not being ready yet to constitute a region of interest.
- the previously described plurality of standard image pixels 91 is 9, and they are equally distributed in an image signal, as shown in FIG. 9B .
- the image pixel and the standard image pixel each has a motion feature value, which is calculated according to the following steps:
- the image signal has the image pixel and the standard image pixel
- the motion parameter of the image pixel and the standard image pixel will need to go through a procedure in which the motion parameter are compared first in terms of discrepancy of location with these image pixels (standard image pixels) in the image signal and the previous image signal in order to obtain an X axis vector MV x of a motion vector, and a Y axis vector MV y of a motion vector.
- the result of such which will need to go through a magnitude operation before a motion parameter is finalized.
- the median number is configured to be the general motion parameter (GMP) of the image signal.
- the motion parameters of the image pixel are configured to be a pixel motion parameter (PMP).
- the previously described step of “individually calculating the motion feature value of the image pixel according to a result from comparing the pixel motion parameter of the image pixel and the general motion parameter of the image signal” is directed to capturing the larger of between the image pixel's pixel motion parameter (PMP) and the image signal's general motion parameter (GMP) to be the motion feature value of the image pixel.
- PMP image pixel's pixel motion parameter
- GMP general motion parameter
- the aforementioned step of “individually calculating the motion feature value according to a result from comparing the pixel motion parameter of the standard image pixel and the general motion parameter of the image signal” is directed to capturing the larger of between the image pixel's pixel motion parameter (PMP) of the standard image pixel and the general motion parameter (GMP) of the image signal.
- PMP pixel motion parameter
- GMP general motion parameter
- each image pixel and standard image pixel is categorized into a plurality of groups according to the motion feature value of the image pixel and the motion feature value of the standard image pixel, which is step (E3).
- the categorizing procedure can include K-means cluster algorithm, CRLA (constraint run length algorithm), or SBKM (symmetry distance based K-means algorithm).
- the preferred algorithm is K-means algorithm.
- the next immediate step will be to capture a group for constituting a region of interest according to a result from categorizing the standard image pixels into the plurality of groups, which is the result of categorizing the 9 standard image pixels into a plurality of groups.
- these 9 standard image pixels are categorized into 2 groups, in which one group includes a motion feature value being equal to a motion feature value of lowest threshold value of a plurality of standard image pixels, and another group includes a motion feature value being larger than the lowest threshold value of a plurality of standard image pixels.
- the lowest threshold for motion feature value is 4.
- the plurality of image pixels of the image signal is also categorized into the aforesaid 2 groups, in which one group includes a motion feature value being equal to a motion feature value of lowest threshold value of a plurality of image pixels, and another group includes a motion feature value being larger than the lowest threshold value of a plurality of image pixels.
- the following step is to capture a group having lesser number of standard image pixels to form a group of region of interest, meaning that the number of standard image pixels of the group for constituting a region of interest is smaller than the number of standard image pixels of the group for not being ready yet to constitute a region of interest.
- the captured group for constituting a region of interest includes a plurality of image pixels and a plurality of standard image pixels, which are categorized into the plurality image pixels and the plurality of standard image pixels of the group of region of interest, which are configured to be the region of interest of the image signal, which is step (E4).
- step (A) of the three-dimensional image content generation method of the present invention as shown in FIG. 9D .
- there are in total 5 texture color feature regions located in FIG. 9D and they are each a first texture color feature region 92 , a second texture color feature region 93 , a third texture color feature region 94 , a fourth texture color feature region 95 and a fifth texture color feature region 96 .
- the range enclosed by the broken line is the region of interest 97 .
- the next step is to execute a overlapping operation algorithm process for the depth map generation process of a region of interest, for the three-dimensional image content generation method of an embodiment of the present invention, so as to individually calculate the overlapping operation for each texture color feature region.
- the overlapping operations of the remaining 4 texture color feature regions are higher than one threshold value for overlapping operation.
- the overlapping operation threshold value is between 50% and 80%.
- the overlapping operation threshold value for the present embodiment is 60%.
- region of interest is directed to the whole two-dimensional image signal, especially the portion that requires extra attention, including the portion of texture color feature region corresponding to a moving object.
- aspect ratio test process is used in calculating aspect ratio value for each texture color feature region, and the overlapping percentage
- the aspect ratio value and the overlapping percentage for each texture color feature region will have been calculated. Then according to the aspect ratio value and aspect ratio threshold value of each texture color feature region, and the overlapping percentage and overlapping percentage threshold, a depth value is designated to each texture color feature region. A depth map corresponding to the image signal can thereby be produced.
- the image region distribution map includes 5 texture color feature regions, each corresponds to the first texture color feature region 101 of a user's head, the second texture color feature region 102 of a user's head, the third texture color feature region 103 of a user's head, the fourth texture color feature region 104 of a user's head, and a fifth texture color feature region for the background.
- step (D2) of the depth map generation graph of FIG. 8 an aspect ratio test process and an overlapping percentage algorithm process are performed on the 5 texture color feature regions.
- the aforementioned aspect ratio test process points to executing an aspect ratio capture method for a texture color feature region.
- the aspect ratio capture method includes the following steps:
- a rectangular frame encloses within it the texture color feature region 106 , and making the rectangular frame and the edge of the texture color feature region 106 fit squarely.
- the next step is to adjust the size and location of the rectangular frame in the two-dimensional image signal, making the adjusted rectangular frame enclose 70% of the area of the texture color feature region, as shown in FIG. 10D .
- the aspect ratio threshold value is between 1.1 and 1.3, where 1.1765 or 1.25 is preferred.
- the aspect ratio value (0.45) of the first texture color feature region 101 is lower than the aforesaid aspect ratio threshold value, the depth value designated to the first texture color feature region 101 is identical to another texture color feature region neighboring and perpendicular to the former.
- texture color feature regions execute aforesaid aspect ratio test process, and it is found that it corresponds to a standing object, for example as aforementioned the first texture color feature region 101 of a user's head, the second texture color feature region 102 of a user's hair, the third texture color feature region 103 of a user's fingers and the fourth color feature region 104 of a user's body all have been determined to correspond to a human user, these texture color feature regions can then be reasonably designated equivalent depth value (with respect to the background behind the user, the depth value of the texture color feature region belonging to the same user can be considered to be entirely equivalent).
- the depth map generation process of the three-dimensional image content generation method of an embodiment of the present invention discloses how to designate a depth value to a texture color feature region.
- the depth value designated to the texture color feature region can be calculated by the following formula:
- VLP is the location of the lowest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region.
- h is a height value of the two-dimensional image signal.
- FIG. 11A shows a distribution map for the texture color feature regions in the two-dimensional image signal.
- the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, or the overlapping percentage of the texture color feature region is higher than the overlapping percentage threshold value
- the depth values of the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region are all of a value represented by D.
- the depth value D designated to the texture color feature region can be determined by the following formula:
- VHP is the location of the highest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region
- VLP is the location of the lowest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region
- VCP is the location of an image pixel to be assigned a depth value in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region
- h is the height value of the two-dimensional image signal
- D min is the smallest depth value, the value of which is
- D max is the largest depth value, the value which is
- FIG. 11B demonstrates the distribution state of the texture color feature regions in the two-dimensional image signal.
- the depth value of an image pixel of which the depth value is to be determined later in a texture color feature region corresponds to an image pixel by a linear relationship in the two-dimensional image signal. Such relationship is demonstrated in FIG. 11C .
- the aforementioned linear relationship can also work to alter other types of relationship, such as second order or third order relationship.
- the aforementioned depth map generation method can also produce an adjusted depth map corresponding to an image signal.
- the depth map generation method includes the following steps:
- steps (F1) and (F2) are identical to the steps (D1) and (D2) of the depth map generation method of FIG. 8 , the detailed operation process of steps (F1) and (F2) and any possible types will not be further elaborated here.
- step (F3) The detailed operation process of step (F3) will, however, be provided below:
- step (F3) which comprises:
- the depth map based on an earlier time point works according to a depth map operation formula based on an earlier time point, according to a forward motion vector and a backward motion vector, and from the depth map the result is calculated.
- the depth map operation formula based on an earlier time point is:
- x ′ x - ( FW - MV x - BW - MV x ⁇ ) 2 ) ( Formula ⁇ ⁇ 16 )
- y ′ y - ( FW - MV y - BW - MV y ⁇ ) 2 ) ( Formula ⁇ ⁇ 17 )
- DVT(x,y,t) is the location of the texture color feature region in the depth map in the two-dimensional image signal.
- DVT (x′,y′,t ⁇ 1) is the location of the texture color feature region in a depth map based on an earlier time point
- FW_MV x is an x-axis vector of a forward motion vector
- FW_MV y is a y-axis vector of the forward motion vector
- BW_MV x is an x-axis vector of a backward motion vector
- BW_MV y is a y-axis vector of the backward motion vector.
- FIG. 13B the relationship of each parameters of the aforementioned Formula 15, Formula 16, Formula 17 can be observed in FIG. 13B .
- the depth map designated to the texture color feature region in the depth map and the depth map based on an earlier time point, designated to the texture color feature region will be compared against each other.
- individually designating a depth value to each texture color feature region As such, an adjusted depth map will be produced. As such, an adjusted depth map will be produced. Also, depending on a predetermined setting in the aforementioned depth map generation method, individually designating an adjusted depth value to each texture color feature region according to a depth value of a texture color feature region in the depth map, and a depth value of a texture color feature region in the depth map based on an earlier time point.
- a depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point and a depth value (DVS) designated to the same texture color feature region in a depth map is lower than an adjusted threshold value ( ⁇ E), (which is DVT ⁇ DVS ⁇ E); and (2) when the depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point is lower than the result of 1 subtracting the adjusted threshold value ( ⁇ E), (which is DVT ⁇ 1 ⁇ E), the depth value D designated to the texture color feature region in the adjusted depth map is to be set as:
- the depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, plus the adjusted threshold value ( ⁇ E), which is represented by D DVT+ ⁇ E.
- the depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, minus the adjusted threshold value ( ⁇ E), which is represented by D DVT ⁇ E.
- the depth value D designated to the texture color feature region in the adjusted depth map. is:
- the depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, which is represented by D DVT
- the three-dimensional image generation process of the three-dimensional image content generation method of the present invention includes the following steps:
- H2 individually calculating a two-dimensional image signal corresponding to the left eyesight and a two-dimensional image signal corresponding to the right eyesight, according to the location of the texture color feature region in the image region distribution map, the depth value designated to the texture color feature region in the depth map.
- the next step involves individually calculating a two-dimensional image signal corresponding to the left eyesight and a two-dimensional image signal corresponding to the right eyesight according to the difference value between the two eyesights of a viewer.
- the algorithm for the above purpose is commonly known to persons of ordinary skills in the art, therefore elaboration for such art will not be provided here.
- a frame rate will be used to individually chain the plurality of three-dimensional image signals responsive to different time point together to form a three-dimensional image signal (three-dimensional image steam).
- the aforementioned frame rate is directed to the number of image signals played per second, and the value of which has been confirmed before being executed by the aforementioned image signal chaining process.
- a video signal has a plurality of image signals, but in some specific applications, this arrangement can also include a video signal.
- frame rate of 24 (which means playing 24 image signals per second) is sufficient for numerous application situations, for example such may be used to display an object moving at a regular speed inside a screen monitor.
- the frame rate of a video signal should at least be 30, or above 40, in order to clearly display the graphical details of a high speed racing car.
- image signals for example three-dimensional image signals
- video signal for example three-dimensional video signal
- the three-dimensional image content generation method of the present invention is understood to be capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
- Processing Or Creating Images (AREA)
Abstract
Description
- This application claims the benefit of filing date of U.S. Provisional Application Ser. No. 61/478,965, entitled “Method for Converting from a 2-Dimensional Video to a 3-Dimensional Video” filed Apr. 26, 2011 under 35 USC §119(e)(1).
- 1. Field of the Invention
- The present invention relates to a method for generating three-dimensional video content particularly to a method that is capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- 2. Description of Related Art
- Growing popularity for 3D motion capture filming techniques has set an inspiring tone for the global community. Added interest to this hotly discussed topic for scientific research about conversion from a two-dimensional image to three-dimensional image was sparkled by the release of the 3D film Avatar directed by the legendary film director James Cameron. A great deal of efforts has been invested to pursue and develop a more advanced image conversion technique aiming for converting two-dimensionality to three-dimensionality in image processing. As well known to persons of ordinary skill in the relevant art, the two main and commonly adapted 3D imaging technique may be categorized into two methods: one being a array-based filming technique, the other being a depth-based filming technique.
- It is known that array-based filming technique involves setting a number of video cameras in an array arrangement, where its general practice involves using two video cameras. However, these two video cameras should undergo a corrective process before filming starts so as to ensure that there is no horizontal position displacement issue between these two video cameras. Furthermore the distance between these two video cameras should emulate the distance between a normal person's eyes (i.e. the visual parallax of the eyes), so as to simulate a possible perspective of a normal person. In another aspect, the requirement for these two video cameras to be maintained in a synchronous state at a run-time setting incurs extraneous effort and time during use of the equipments. In another aspect, these two video cameras may be subject to changes in light abundance or changes in internal parameters of the video cameras, making the three-dimensional images so captured open to discoloring.
- It is also known that depth-based filming technique turns to use one traditional video camera while also coupling with another depth video camera for filming purposes. Generally speaking, the aforementioned depth video camera operates on a “time of flight” principle, working by means of calculating the time period from the emitted infrared light hitting the desired object to the same infrared light returning back into the depth video camera, so as to determine the actual distance between the desired object for phototaking and the video camera. In addition, the relative distance between the aforementioned traditional video camera and the aforementioned depth video camera may be subject to adjustment prior to filming begins so as to ensure that the pixels picked up by the traditional video camera are all well aligned with their corresponding depth value. Unfortunately, the depth resolution offered by the depth video camera still faces serious limitation, and it is not possible to provide a depth image of high image resolution, which altogether creates room for improving the image quality of composite three-dimensional image.
- Inasmuch the aforementioned two techniques may already be capable of forming three-dimensional image content, the underlying issue of increased operating cost with these setups remain to be of concern as, for example, the array-based filming technique requires a multiple number of video cameras to be on standby during filming. Taking into another consideration, another disadvantage hereof is that the three-dimensional image produced therefrom will be more open to distortion and become useless if each participating video camera is not adjusted relative to its position from one another or does not synchronize accordingly. Another disadvantage associated with the known setups is of the high price for a depth video camera, which has demonstrated to be a barrier for higher market penetration, making it less possible to effectively lower operating cost for producing a three-dimensional video content with a depth video camera.
- Owning to the disadvantage of high filming cost arising from three-dimensional image capturing, effective approach for maintaining quality image while keeping the rate low on producing three-dimensional images has been a popular issue for further research in the industry and academia. Also, two-dimensional image signal is the type of image signal that is most commonly used in the current relevant state of art. It will therefore be understood that the issues encountered by the aforementioned ways for obtaining three-dimensional image content (three-dimensional image signal) will be overcome if a new way for converting two-dimensionality into three-dimensionality is found.
- Accordingly, there is a demand in the industrial sectors for a method for receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- One object of the present invention is to provide a method for generating three-dimensional video content particularly to a method that is capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- According to one aspect of the present invention, there is provided a method for generating a three-dimensional image signal, comprising the following steps: (A) receiving a 2D image and generating a region of interest distribution map from the 2D image, wherein the two-dimensional image signal comprises a plurality of individually colored image pixel, and the region of interest distribution map comprises a region of interest; (B) executing a color feature capture process, for forming a plurality of color feature regions; (C) executing an image segmentation process basing on the texture feature of the plurality of color feature regions, for forming an image region distribution map; (D) executing a depth map generation process, for generating a depth map basing on the region of interest distribution map and the image region distribution map, wherein in the depth map each color texture feature region is individually characterized by a depth value; (E) executing a 3D image generation process, for forming the 3D image basing on the image region distribution map and the depth map; and (F) chaining a plurality of the 3D images to form the 3D video basing on a frame rate.
- It is therefore apparent through a three-dimensional image content generation method disclosed by the present invention to receive a two-dimensional image signal and direct the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
-
FIG. 1 illustrates a flow chart of steps during operation of the three-dimensional image content generation method according to an embodiment of the present invention; -
FIG. 2 illustrates a flow chart of steps during operation of the color capture process of the three-dimensional image content generation method according to an embodiment of the present invention; -
FIG. 3A is a schematic view illustrating the process of respectively mapping the colors of two image pixels contained in a two-dimensional image signal to a color plane of an HSI color space, for respectively clustering the colors of the two image pixels into a color region or a fuzzy region; -
FIG. 3B is a graph showing the result from the clustring as illustrated inFIG. 3A according to the present invention with respect to the process of designating the first image pixel with a color feature vector; -
FIG. 3C is a graph showing the result from the clustring as illustrated inFIG. 3A according to the present invention with respect to the process of designating the second image pixel with a color feature vector; -
FIG. 4 illustrates a flowchart of steps during operation of an image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention; -
FIG. 5 is a diagram illustrating the disposition of twelve two-dimensional Gabor filters included in a Gabor filter package on a spatial plane, operational in the image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention; -
FIG. 6A shows a distribution diagram for a plurality of color feature regions represented in a two-dimensional image signal; -
FIG. 6B shows a distribution diagram for a plurality of color feature regions ofFIG. 6A as further segmentized after a re-segmentation process; -
FIG. 7A illustrates a flow chart of steps during operation of an image segment merging process according to an embodiment of the present invention; -
FIG. 7B is a distribution diagram for a plurality of texture color feature regions represented in a two-dimensional image signal; -
FIG. 7C is an illustrative diagram demonstrating the result of merging a texture color feature regions oftag number 1 and a texture color feature regions oftag number 5 for deriving a transitorily merged texture color feature regions; -
FIG. 8 illustrates a flow chart of steps during operation of a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention; -
FIG. 9A illustrates a flow chart of steps during operation of generating a graph showing distribution for region of interest of the three-dimensional image content generation method according to an embodiment of the present invention; -
FIG. 9B is an illustrative diagram showing distribution of nine default image pixels of an image signal; -
FIG. 9C shows a flow chart of steps during operation of calculating for each of a plurality of image pixels and motion feature values of a plurality of default image pixels; -
FIG. 9D is a distribution diagram for five texture color feature regions and one region of interest in a two-dimensional image signal; -
FIG. 9E is an illustrative diagram exhibiting the result of designating equal depth values among four texture color feature regions according to the five texture color feature regions ofFIG. 9D ; -
FIG. 10A is a diagram showing distribution of a plurality of texture color feature regions of an image region; -
FIG. 10B is a flow chart of steps during operation of an aspect ratio value capture procedure required for a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention; -
FIG. 10C is a diagram demonstrating a rectangular frame enclosing a portion of the texture color feature region, and cropping the edges of the texture color feature region; -
FIG. 10D is a diagram showing an adjusted rectangular frame enclosing a 70% area of the texture color feature region as presented after the size and position of the rectangular frame onto a two-dimensional image signal is adjusted; -
FIG. 11A is a diagram displaying designation of an identical depth value for a plurality of image pixels; -
FIG. 11B is a diagram displaying designation of a linear distribution of depth values for a plurality of image pixels; -
FIG. 11C shows a chart exhibiting the linear relationship corresponding to a depth value for use in attributing a pending depth value to an image pixel; -
FIG. 12 is another embodiment for a flow chart of steps during operation of depth map generation process for an adjusted depth map operable to produce a corresponding image signal; -
FIG. 13A shows a chart of sub-steps for the step F3 of the depth map generation process according to an embodiment ofFIG. 12 ; -
FIG. 13B is a diagram showing the relationship between each of the parameters used in generating an anterior depth map based on a forward motion vector and a backward motion vector; and -
FIG. 14 illustrates a flow chart of steps during operation of the three-dimensional image formulation process of the three-dimensional image content generation method according to an embodiment of the present invention. -
- 31 Red color region
- 32 Green color region
- 33 Cyan color region
- 34 Blue color region
- 35 Red-violet color region
- 51 Internal subband Gabor filter
- 52 External subband Gabor filter
- 91, 101 First color texture feature region
- 92, 102 Second color texture feature region
- 93, 103 Third color texture feature region
- 94, 104 Fourth color texture feature region
- 95, 105 Fifth color texture feature region
- 97 Region of interest
- 106 Color texture feature region
- 107 70% area coverage of color texture feature region
- The invention is capable of other embodiments and of being practiced or carried out in various ways. Variations and modifications of the foregoing are within the scope of the present invention. It will also be understood that the invention disclosed and defined herein extends to all alternative combinations of two or more of the individual features mentioned or evident from the text and/or drawings. All of these different combinations constitute various alternative aspects of the present invention. The embodiments described herein explain the best modes known for practicing the invention and will enable others skilled in the art to utilize the invention.
-
FIG. 1 shows a flow chart of steps during operation of the three-dimensional image content generation method according to an embodiment of the present invention. An advantage of the three-dimensional image content generation method may be for transforming a two-dimensional image signal into a three-dimensional image signal. As shown inFIG. 1 , the method for generating a three-dimensional image content comprises the following steps: - (A) receiving a 2D image and generating a ROI distribution map from the 2D image;
- (B) executing a color feature capture process, for forming a plural of color feature regions;
- (C) executing an image segmentation process basing on the texture feature of the plural of color feature regions, for forming an image region distribution map;
- (D) executing a depth map generation process, for generating a depth map basing on the ROI distribution map and the image region distribution map;
- (E) executing a 3D image generation process, for forming the 3D image basing on the image region distribution map and the depth map; and
- (F) chaining a plurality of the 3D images to form the 3D video basing on a frame rate.
- Step (B) of the three-dimensional image content generation method of the present invention may be referred to
FIG. 2 , which is directed to a flow chart of steps during operation of the color capture process of the three-dimensional image content generation method according to an embodiment of the present invention, wherein the steps of which comprise the following: - (A1) receiving the two-dimensional image signal;
- (A2) mapping the image signal to a color space model, enabling the color contained by each of the plurality of image pixels to be respectively mapped to a corresponding point in the color space model, the corresponding point having a coordinate represented by a first parameter, a second parameter, and a third parameter, the color space model including a color plane composed of a plurality of color regions and a plurality of fuzzy regions;
- (A3) comparing the third parameter with an adjustment threshold value to work out a comparison result, and obtaining an adjusted second parameter based on the comparison result and the second parameter;
- (A4) sequentially clustering each of the color provided by the plurality of image pixels into one of the plurality of color regions or one of the plural fuzzy regions of the color plane according to the first parameter and the adjusted second parameter;
- (A5) designating a color feature vector to each of the plurality of image pixels according to a result of clustering;
- (A6) applying a clustering algorithm, and clustering the image pixels into a plurality of color clusters based on the color feature vector designated respectively to each of the image pixel;
- (A7) encircling in an identical color feature region the image pixels that border each other and belong in same color group.
- In step (A1) of the abovementioned method, one two-dimensional image signal includes a plurality of image pixels, and these image pixels each contains a color. In addition, the two-dimensional image signal format is unlimited, which may be YcbCr420 or RGB444. Next, in step (A2), the image signal is mapped to a color space model, making the colors contained by each image pixels projected individually to a pair of corresponding points in the color space model, and the coordination for the corresponding point is composed of a first parameter value, a second parameter value, and a third parameter value. In the present embodiment, the color space model may be of any desirable type, which may be a HSI color space model, a RGB color space model, a YUV color space model or a CMY color space model.
- Referring now to
FIG. 3A , a schematic view is provided illustrating the process of respectively mapping the colors of two image pixels contained in a two-dimensional image signal to a color plane of an HSI color space, for respectively clustering the colors of the two image pixels into a color region or a fuzzy region. - As shown in
FIG. 3A , the aforementioned color feature capture process works to divide a color plane of the HIS color space model into five color regions, each representing a first color region, a second color region, a third color region, a fourth color region and a fifth color region, wherein each color region individually responds to a particular color. In the present embodiment, the first color region corresponds to a red region 31 (red), the second color region corresponds to a green region 32 (green), the third color segment corresponds to a cyan region 33 (cyan), the fourth color region corresponds to a blue region 34 (blue), and the fifth color segment corresponds to a magenta region 35 (magenta). - In addition, as can be seen on the color plane of
FIG. 3A , the contained angle inred region 31 is between 350° to 410°, the contained angle ingreen region 32 is between 95° to 120°, the contained angle incyan region 33 between 170° to 190°, the contained angle inblue region 34 is between 230° to 250°, the contained angle inmagenta region 35 is between 290° to 310°. It is to noted here that the aforementioned contained angles all can not only include the above five angle ranges, but also be adjusted with respect to actual demand accordingly. Furthermore, a fuzzy region is defined as a space located between two neighboring color regions (those not covered by the five color regions as mentioned above). For example, a region located between a first color region and a second color region, or a region located between a third color region and a fourth color region. Therefore, as shown byFIG. 3A , the color plane has five color regions and five fuzzy regions. - In the present embodiment, the two image pixels contained by the aforementioned image signal is each a first image pixel and a second image pixel, and these two image pixels each has a color, as previously recited by step (A1), the color feature capture process.
- In another embodiment of the present invention, as described by step (A2) for color feature capture process, the aforementioned image signal is mapped to a HSI color space model, making the colors contained by first image pixel and second image pixel mapped to a corresponding point in the HSI color space model. In the present case, the corresponding point P1 for the color of the first image pixel in the HSI color space model is (105°, 0.75, 90), the corresponding point P2 for the color of the second image pixel in the HSI color space model is (130°, 0.625, 85), wherein the first parameter is for a hue value, the second parameter is a saturation value, the third parameter is an intensity value.
- As described by step (A3) for color feature capture process, the next step is to compare the third parameter value of the corresponding point P1 and P2, which is the intensity value, with an adjustment threshold value, and then work out an adjusted second parameter value based on a result of the comparison and their respective second parameter value, which is again, saturation value. In a preferred embodiment of the present invention, an adjusted threshold value is between 70 and 90, in an even more preferred embodiment, the value is between 80 and 87. In the current embodiment, the aforementioned adjusted threshold value is 85. When the intensity value of either of the corresponding point P1 or corresponding point P2 is higher than the aforementioned adjusted threshold value, the following formula will be used to calculate the adjusted saturation component:
-
S′=a+b*S (Formula 1) - wherein S′ is an adjusted saturation component, S is a saturation component, a and b are each a random positive number, and satisfy the a+b=1 condition. In the present example, the abovementioned formula can further modified as:
-
S′=0.2+0.8*S (Formula 2) - wherein a=0.2, b=0.8. However, the values of a and b may change for other embodiments; for example, a=0.3, b=0.7.
- In another situation, when the intensity value of either of the corresponding point P1 or corresponding point P2 is not higher than the aforementioned adjusted threshold value, the following formula will be used to calculate the adjusted saturation component:
-
- wherein S′ is an adjusted saturation value, S is saturation value, T is the adjusted threshold value, a and b is each a random positive number, and satisfy the a+b=1 condition. In the present example, the above-mentioned formula can be further modified as:
-
- wherein a=0.2, b=0.8. However, the values of a and b may change for other embodiments; for example, a=0.3, b=0.7. Similarly, although the adjusted threshold value T is 85, the adjusted threshold value T may also be other values, such as 90 or 80.
- As described above, in the present example, the corresponding point P1 for the color of the first image pixel in the HSI color space model is (105°, 0.75, 90), and the corresponding point P2 for the color of the second image pixel in the HSI color space model is (130°, 0.625, 85). In this case, since the intensity value (90) of P1 is higher than the adjusted threshold value (85), the adjusted saturation value will work out to be 0.8 as calculated by the
aforementioned formula 2. As a result, the coordination of P1 inFIG. 3A on a color plane is (105°, 0.8). In another situation, since the intensity value (85) of P2 is not higher than the aforementioned adjusted threshold value (85), the adjusted saturation value will work out to be 0.8 as calculated by theaforementioned formula 4. As a result, the coordination of P2 inFIG. 2A on a color plane is (130°, 0.7). - In step (A4) of the aforementioned color feature capture process, the colors contained by two image pixels are categorized into one of the five color regions or one of the five fuzzy regions of a color plane as shown in
FIG. 3A , wherein the two image pixels are determined from the first parameter and the adjusted second parameter of the corresponding point P1 and P2 respectively, which in other terms may be referred to as the hue value and the adjusted saturation value. - Referring now to
FIG. 3A , the corresponding point P1 is located at (105°, 0.8), and can be categorized into thegreen region 32, and because the corresponding point P2 is located at (130°, 0.7), it is categorized into thefuzzy region 36 between thegreen region 32 andcyan region 33. - Next following the color feature capture process as described by step (A5) of the above, a color feature vector is designated to each of these image pixels according to a result of the abovementioned categorization. When the color of the image pixel is classified into one of the five color regions, the color feature vectors designated to the image pixel include a first vector, a second vector, a third vector, a fourth vector, a fifth vector, and a sixth vector. Furthermore, in the present embodiment the first vector responds to a
red region 31, the second vector responds to agreen region 32, the third vector responds to acyan region 33, the fourth vector responds to ablue region 34, the fifth vector responds to amagenta region 35, and the sixth vector responds to an achromatic region. - In addition, the value of the sixth component is equal to 1 subtracting adjusted saturation value (i.e. 1−S′), while also the sum of the values for the first vector, second vector, third vector, fourth vector, fifth vector, and sixth vector is 1. In the present example, only the component of which the color of the corresponding image pixel is categorized to the color region (i.e. the second vector), and the sixth vector are of non-zero values, and the sum of the two values is 1. In the situation just described, only the vector of which the color of the corresponding image pixel is categorized to the color region (i.e. the second component), and the sixth vector are of non-zero values, and the sum of the two values is 1.
- Generally speaking, the achromatic color feature relates to the amount of white light component within a color of an image pixel. The amount of such component can enable a viewer see the difference between a color's saturation level. More specifically, human perception to an image pixel may differ significantly if the colors of which share the same hue but different level of saturation. For example, the hue components of bright red or dark red may be very similar, but differ significantly with respect to saturation; this is because the white light components differ by a significant degree, which therefore has stimulated various sources defining achromatic color feature as for aiding in capturing features having improved color properties.
- However, when the color of an image pixel is not categorized into the aforementioned five color regions, for example being categorized into one of the five fuzzy regions, the color feature vectors designated to the image pixel will include a first vector, a second vector, a third vector, a fourth vector, a fifth vector, and a sixth vector. To recite from before, the value of the sixth vector is equal to 1 minus the adjusted saturation component, and its mathematical representation may be 1−S′. Moreover, the sum of the values for the first vector, second vector, third vector, fourth vector, fifth vector, and the sixth vector is 1. In the present example, only two vectors that respectively correspond to the color regions on the two sides of the aforementioned fuzzy region (i.e. the second vector and the third vector), and the sixth vector are of non-zero values, and the sum of the three values is 1.
- Since the corresponding point P1 is categorized into the
green region 32, it will be understood that only the second vector and the sixth vector corresponding to thegreen region 32 show non-zero values, as shown inFIG. 3B . Even more, since the adjusted saturation value for P1 is 0.8, the second vector corresponding to thegreen region 32 in the aforementioned color feature vector is 0.8, and the value of the sixth vector is 0.8 (since 1−0.8=0.2). Accordingly, it will be understood to designate a color feature vector, represented by V1(0,0.8,0,0,0,0.2). - In another perspective of the invention, since P2 is categorized into the
fuzzy region 36 between thegreen region 32 andcyan region 33, it will also be understood that only the second vector corresponding to thegreen region 32, the third vector corresponding to thecyan region 33, and the sixth vector will show non-zero values. In addition to this, if the color contained by the image pixel is not categorized into one of the aforementioned five color regions, and P2 is categorized into a fuzzy region between agreen region 32 and acyan region 33, the value of the second vector may be calculated using the following formula: -
C=S′*Belongingness (Formula 5) - where C is a value of the second vector, S′ is the adjusted saturation component, Belongingness is a degree of belongingness, and the degree of belongingness may be determined by:
-
- where U is an angle contained by a distant border of a color region on the color plane, to which the second vector corresponds, L is an angle contained between a neighboring border of a color region on the color plane, to which the second vector corresponds, H is an angle contained between the hue component and the color plane.
- Taking P2 as an example, since P2 is categorized into a fuzzy region between a
green region 32 and acyan region 33, it is required to individually determine the fuzzy region's belongingness with respect to its two neighboring color regions (green region 32 and cyan region 33). The first step relates to thegreen region 32, and because P2 is located at (130°, 0.7), the angle that its hue component contains on the color plane is 130°. Next, the fuzzy region has two borders, and one distant border of thegreen region 32 on the color plane (L) is 120°. Therefore, the fuzzy region's belongingness to thegreen region 32 is: -
- Therefore, given that the adjusted saturation component is 0.7, couple with the
aforementioned formula 5, it can be determined that the vector quantity corresponding togreen region 32 is 0.7*0.80.56, as shown inFIG. 3C . - By the same principle using the same definitions, the required parameters for calculating P2's belongingness with respect to the
cyan region 33 will be: an angle contained by one distant border of thecyan region 33 on the color plane (U) is 120°, and one angle contained by one neighboring border of thecyan region 33 on the color plane (L) is 170°. Accordingly, the belongingness of P2 with respect to thecyan region 33 may be calculated as follows: -
- Therefore, since the adjusted saturation component of P2 is 0.7, and in conjunction with the
aforementioned formula 5, it may be determined that the vector quantity of thecyan region 33 is 0.7*0.2=0.14, as shown inFIG. 3C . Also since the adjusted saturation component is 0.7, the sixth component is 0.3 (1−0.7=0.3). Then, it will be understood that an image pixel's color may be used to designate an image pixel's color feature vector V2(0,0.56, 0.14, 0, 0, 0.3). - Subsequently in step (A6) of the above process, a clustering algorithm is used to cluster image pixels into clusters based on the designated color feature vector of the image pixel. In general, the cluster algorithm herein may be K-means cluster algorithm methodology, CRLA (constraint run length algorithm) cluster algorithm methodology, SBKM (symmetry distance based K-means algorithm) cluster algorithm methodology, etc., but the preferred methodology is K-means cluster algorithm.
- Lastly, the color feature capture process of step (B) of the three-dimensional image content generation method of the present invention is capable of precisely capturing a color feature vector from the colors of a two-dimensional image signal having a plurality of image pixels, such as V1(0,0.8,0,0,0,0.2) and V2(0,0.56,0.14,0,0,0.3), and then designating these two color feature vectors to their corresponding image pixels. Next, by putting into application of a cluster algorithm, the image pixels of a two-dimensional image signal can be categorized into a plurality of color clusters based on the color feature vector designated to each image pixel. Finally, a plurality of neighboring and being-of-same-color-cluster image pixels are included in the same color feature region, thereby making it possible to form a plurality of color feature regions.
- Next, please refer to
FIG. 4 for the step (C) for an image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention.FIG. 4 illustrates a flowchart of steps during operation of an image segmentation process using texture feature of the three-dimensional image content generation method, according to an embodiment of the present invention, wherefore the process includes the following steps: - (B1) receiving the two-dimensional image signal;
- (B2) applying a Gabor filter, executing a Gabor filtering process on each image signal, and executing further a value algorithm process on a result from the Gabor filtering process;
- (B3) designating individually a texture feature vector to the image pixel, determined based on a result of the value algorithm process;
- (B4) executing a segmentation process on the two-dimensional image pixels, based on the designated texture feature vector of the image pixel such that the two-dimensional image signal has a plurality of texture feature segments, and such that the image pixels located within the same texture feature region each has texture feature vector of identical cluster;
- (B5) individually executing a second segmentation process in accordance with distribution of the texture feature region in the two-dimensional image signal, such that at least one color feature region has a plurality of texture color feature region, and such that the image pixels located in the same texture color feature region each has texture feature vector of identical cluster and color feature vector of identical cluster; and
- (B6) forming an image region distribution map based on the texture color feature region in the two-dimensional image signal.
- Firstly, as per step (B1), the two-dimensional image signal comprises a plurality of image pixels. There is no limitation on the choice of format for the two-dimensional image signal, and such lack of limitation allows a two-dimensional image signal represented by any format to be used in the image segmentation process using texture feature.
- Next, in step (B2), a Gabor filter is installed with the two-dimensional image signal to implement a Gabor filtering process to each image pixel, and subsequently execute a value algorithm process with respect to the output from the execution of the Gabor filtering process. In further illustration, such value algorithm process is for calculating the modulation magnitude outputted from the Gabor filtering process. In addition in the present example, the Gabor filter package includes an internal
subband Gabor filter 51 and an externalsubband Gabor filter 52, and the internalsubband Gabor filter 51 and externalsubband Gabor filter 52 each includes six two-dimensional Gabor filters, meaning that each of the internalsubband Gabor filter 51 and the externalsubband Gabor filter 52 has an identical number of two-dimensional Gabor filters. A schematic view of the distribution of the twelve two-dimensional Gabor filters on the space plane is illustrated inFIG. 5 . - Furthermore, the twelve two-dimensional Gabor filters can each be described by the following formula:
-
- where each σx, σy each is a Gaussian kernel standard deviation, ω is a Fourier fundamental frequency, θ is a direction of the two-dimensional Gabor filter. Also, the relationship between x′, y′, x′, y satisfy the following formulas:
-
x′=x cos θ+y sin θ (Formula 10) -
y′=x sin θ+y cos θ (Formula 11) - In the present example, the standard deviations for the six two-dimensional Gabor filters of the internal
subband Gabor filter 51, σx, σy are preferred to be configured at between 1.75 and 6.5, and the standard deviations for the six two-dimensional Gabor filters of the internalsubband Gabor filter 51, σx, σy are preferred to be configured at between 1.75 and 4.5. In addition, the Fourier fundamental frequency ω is preferred to be configured at between 0.2π and 0.6π. - The next step (B3) involves individually designating each of the image pixel of the two-dimensional image signal a texture feature vector in accordance with the output (modulation magnitude) obtained from executing the aforementioned value algorithm process. The number of dimensions of these texture feature vectors is equal to the number of the two-dimensional Gabor filters used in the Gabor filtering process in step (B2).
- In step (B4), a segmentation process is implemented on the two-dimensional image signal according to the texture feature vector designated to each image pixel, such that the two-dimensional image signal has a plurality of texture feature regions. Additionally, the plurality of image pixels in the same texture feature region each has a texture feature vector of identical cluster. Because of this, the two-dimensional image signals that are results of a segmentation process are segmented into a plurality of texture color feature region, and these texture color feature regions can be distributed to any location within the entire area realized contained in the two-dimensional image signal. The segmentation process in the present embodiment uses a K-means cluster algorithm. Since K-means cluster algorithm is widely used in numerous fields of technology and is commonly known by people skilled in the relevant art, an elaboration of the K-means cluster algorithm is omitted.
- Step (B5) involves individually executing a second segmentation process in accordance with distribution of the texture feature region in the two-dimensional image signal, such as their relative positions in the two-dimensional image signal, such that at least one color feature region has a plurality of texture color feature region. The second segmentation process herein involves the following steps:
- Comparing with each other the value of the area of the color feature region and the value of the second segmentation process; and
- When the area of one of the color feature region is higher in value than the threshold value after second segmentation process, performing a third segmentation process on the color feature region to achieve the texture color feature region by following the distribution of texture feature regions in the two-dimensional image signal.
- Referring now to
FIGS. 6A and 6B for the result of the aforementioned second segmentation process,FIG. 6A shows the distribution of a plurality of color feature regions in a two-dimensional image signal, andFIG. 6B shows a resulting situation after a second segmentation process is finished, wherein the color feature region ofFIG. 6A is further segmented into a plurality of texture color feature regions. - Lastly, step (B6) involves forming an image region distribution map obtained from step (C) of an embodiment of the three-dimensional image content generation method of the present invention, based on the texture color feature region in the two-dimensional image signal.
- The image segmentation process using texture feature of the three-dimensional image content generation method of an embodiment of the present invention has been described therefrom. Nevertheless, after the image segmentation process of step (C) is finished, step (C) can further include a step of executing an image region combination process subject to actual requirement, such that a plurality of texture color feature region is combined into a combined image texture color feature region.
- Please refer to
FIG. 7A for details on the image region combination process. Of which,FIG. 7A is directed to an image region combination process of an embodiment of a three-dimensional image content generation method, and the image region combination method includes the following steps: - (C1) capturing an area value and a perimeter value of each texture color feature region so as to determine a closeness value of each texture color feature region;
- (C2) according to a closeness value for each texture color feature region, arranging the texture color feature regions from a texture color feature region having highest area value to a texture color feature region having lowest area value;
- (C3) according to an order arranged from the texture color feature region having highest area value to the texture color feature region having lowest area value, executing a combination test process on each texture color feature region so as to form a transitory combined image texture color feature region, the transitory combined image texture color feature region includes the texture color feature region and a neighboring texture color feature region, and capturing an area value and a perimeter value of the transitory combined image texture color feature region so as to determine a closeness value of the transitory combined image texture color feature region; and
- (C4) comparing the closeness value of the transitory combined image texture color feature region and the closeness value of one of the texture color feature region, and when the closeness value of the transitory combined image texture color feature region is higher than the closeness value of one of the texture color feature region, setting the transitory combined image texture color feature region as a combined image texture color feature region.
- Heretofore,
FIG. 7B andFIG. 7C are referred for detailing the aforementioned combination test process. As will be seen,FIG. 7B displays the way a plurality of texture color feature region is disposed in a two-dimensional image signal, andFIG. 7C illustrates combination of a texture color feature region labeled as 1 and another texture color feature region labeled as 5 to give rise to a transitory combined image texture color feature region. - Continuing from the above, the closeness value of a plurality of texture color feature region can be determined using the following formula:
-
- where C is closeness value, A is area value, P is perimeter value. Accordingly, the closeness value of each texture color feature region can be determined.
- Next, in the aforementioned image region combination method, according to the order just described (arranging the texture color feature regions from a texture color feature region having highest area value to a texture color feature region having lowest area value), execute a combination test process on each texture color feature region so as to form a transitory combined image texture color feature region, wherein the transitory combined image texture color feature region includes one of the texture color feature regions and another, neighboring texture color feature region. Afterwards, the area value and perimeter value of the transitory combined texture color feature region are determined and ultimately the closeness value is determined accordingly.
- In the present example, the aforementioned combined test process is executed starting with the texture color feature region labeled as 1. As shown in
FIG. 7B , such arrangement begins with the plurality of texture color feature regions neighboring the texture color feature region labeled as 1, for example, combining the texture color feature region labeled as 1 and the texture color feature region labeled as 5 into a transitory combined image texture color feature region, then, combining the texture color feature region labeled as 1 and the texture color feature region labeled as 4 into another transitory combined image texture color feature region, and the above process continues on further. - However in some applications, a threshold value identification process will be executed before the aforementioned “combination test process” is executed, meaning that the area value and closeness value of each of the texture color feature region are compared respectively with an area threshold value and a closeness threshold value. Also, the aforementioned combination test process is only executed when the area value of texture color feature region resulting from the comparison is lower than the aforementioned area threshold value, and when the closeness value of the texture color feature region is lower than the closeness value of the aforementioned threshold value. In the present embodiment, the area threshold value is between 120 to 240, wherein the preferred embodiment is 180; the closeness threshold value is between 0.001 to 0.008, wherein the preferred embodiment is 0.003, depending on the environmental conditions.
- Then as shown in
FIG. 7C , the texture color feature region labeled as 1 and the texture color feature region labeled as 5 are combined into a transitory combined image texture color feature region, then the area value and perimeter value of the transitory combined image texture color feature region obtained the combination of the texture color feature regions labeled as 1 and 5 can be determined. Accordingly, the closeness value of the transitory combined image texture color feature region can be determined. - The closeness value of the transitory combined image texture color feature region and the closeness value of the transitory combined image texture color feature region labeled as 1 are compared. When the closeness value of the transitory combined image texture color feature region is higher than the closeness value of the texture color feature region labeled as 1, the transitory combined image texture color feature region is set as the combined image texture color feature region. On the contrary, when the closeness of the texture color feature region is not higher than the closeness value of the texture color feature region labeled as 1, the aforementioned process is repeated. An instance illustrating this embodiment is to combine the texture color feature region labeled as 1 and the texture color feature region labeled as 4 to form a transitory combined image texture color feature region, and then capture the area value and perimeter value of another transitory combined image texture color feature region to further determine the closeness value of this transitory combined image texture color feature region.
- When the aforementioned combination test procedure is executed on all the texture color feature regions neighboring the texture color feature region labeled as 1, the previously mentioned ordering method (arranging the texture color feature regions from a texture color feature region having highest area value to a texture color feature region having lowest area value) is replaced with designating the texture color feature region of second largest area value to be the standard texture color feature region for performing the combination test procedure.
- Lastly, when the aforementioned combination test process is executed on all the texture color feature regions (for combining the standard texture color feature region for combination test process), the previously described image region combination process is complete, meaning that for the above two-dimensional image signal, all of the texture color feature regions satisfying the combination conditions of texture color feature region are combined according to the relative relationship there between to form a plurality of combined image texture color feature regions.
- As per the depth map generation process of step (D) of the three-dimensional image content generation method of the present invention, please refer to
FIG. 8 , whereinFIG. 8 relates to a flow chart of steps during operation of a depth map generation process of the three-dimensional image content generation method according to an embodiment of the present invention: - (D1) executing an aspect ratio test process on each texture color feature region and a overlapping rate algorithm process on a region of interest; and
- (D2) individually designating a depth value to the texture color feature region according to the result of the aspect ratio test process and the result of the overlapping rate algorithm process.
- Wherein, the aforementioned aspect ratio test process is used for calculating the aspect ratio value of the texture color feature region, and when the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, the depth value designated to the texture color feature region is equal to the depth value designated to another texture color feature region that is perpendicularly neighboring to the former.
- In another aspect of the present invention, the aforementioned overlapping rate algorithm process for the region of interest is used in calculating a overlapping rate for between a texture color feature region and a region of interest (including the distribution map of region of interest of step (A) of the present invention), and all the texture color feature regions that are neighboring each other and also have overlapping operation that is higher than a overlapping operation threshold value are designated the same depth value.
- Next, the previously described distribution map for region of interest is obtained from the execution of a generation process for a distribution map for region of interest. For the generation process for distribution graph for region of interest, please refer to
FIG. 9A , which is a flow chart of steps during operation of generating a graph showing distribution for region of interest of the three-dimensional image content generation method according to an embodiment of the present invention: - (E1) receiving the image signal, wherein the image signal has a plurality of image pixels and a plurality of standard image pixels;
- (E2) calculating the motion feature value of each of the image pixel and the standard image pixel;
- (E3) categorizing the image pixel and the standard image pixel into a plurality of groups according to the motion feature value of the image pixel and the motion feature value of the standard image pixel; and
- (E4) capturing a group for constituting a region of interest from the plurality of groups according to a result from categorizing the standard image pixels into the plurality of groups.
- Wherein, the number of standard image pixels of the group for constituting a region of interest is smaller than the number of standard image pixels of the group for not being ready yet to constitute a region of interest.
- Wherein, in the present embodiment, the previously described plurality of
standard image pixels 91 is 9, and they are equally distributed in an image signal, as shown inFIG. 9B . - In addition, as shown in
FIG. 9C , the image pixel and the standard image pixel each has a motion feature value, which is calculated according to the following steps: - receiving the image signal, the image signal has the image pixel and the standard image pixel;
- calculating the motion feature value for the image pixel and the standard image pixel;
- calculating a general motion feature value according to the motion parameter of the standard image pixel, and calculating a pixel motion parameter of the image pixel according to the motion parameter of the image pixel; and
- individually calculating the motion feature value of the image pixel according to a restful from comparing the pixel motion parameter of the image pixel and the general motion parameter of the image signal, and individually calculating the motion feature value according to a result from comparing the pixel motion parameter of the standard image pixel and the general motion parameter of the image signal.
- Wherein, the motion parameter of the image pixel and the standard image pixel will need to go through a procedure in which the motion parameter are compared first in terms of discrepancy of location with these image pixels (standard image pixels) in the image signal and the previous image signal in order to obtain an X axis vector MVx of a motion vector, and a Y axis vector MVy of a motion vector. The result of such which will need to go through a magnitude operation before a motion parameter is finalized.
- In another perspective of the invention, after obtaining the motion parameters of the 9
standard image pixels 91, a median number operation will need to be gone through in order to capture the median number of these 9 motion parameters. Afterwards, the median number is configured to be the general motion parameter (GMP) of the image signal. Furthermore, the motion parameters of the image pixel are configured to be a pixel motion parameter (PMP). - In yet another embodiment of the present invention, the previously described step of “individually calculating the motion feature value of the image pixel according to a result from comparing the pixel motion parameter of the image pixel and the general motion parameter of the image signal” is directed to capturing the larger of between the image pixel's pixel motion parameter (PMP) and the image signal's general motion parameter (GMP) to be the motion feature value of the image pixel. Similarly, the aforementioned step of “individually calculating the motion feature value according to a result from comparing the pixel motion parameter of the standard image pixel and the general motion parameter of the image signal” is directed to capturing the larger of between the image pixel's pixel motion parameter (PMP) of the standard image pixel and the general motion parameter (GMP) of the image signal. Wherefore from the obtained image signal, all the image pixels and standard image pixels each have a “motion feature value,” which is step (E2).
- With reference now to
FIG. 9A , when the image signal is obtained, and of which all the image pixels and the standard image pixels each has a motion feature value, each image pixel and standard image pixel is categorized into a plurality of groups according to the motion feature value of the image pixel and the motion feature value of the standard image pixel, which is step (E3). It will be more apparent to persons of relevant skills in the art now that the categorizing procedure can include K-means cluster algorithm, CRLA (constraint run length algorithm), or SBKM (symmetry distance based K-means algorithm). However in the present embodiment, the preferred algorithm is K-means algorithm. - The next immediate step will be to capture a group for constituting a region of interest according to a result from categorizing the standard image pixels into the plurality of groups, which is the result of categorizing the 9 standard image pixels into a plurality of groups. In the present embodiment, these 9 standard image pixels are categorized into 2 groups, in which one group includes a motion feature value being equal to a motion feature value of lowest threshold value of a plurality of standard image pixels, and another group includes a motion feature value being larger than the lowest threshold value of a plurality of standard image pixels. In the present embodiment, the lowest threshold for motion feature value is 4.
- In the meantime, in addition to the aforesaid 9 standard image pixels, the plurality of image pixels of the image signal is also categorized into the aforesaid 2 groups, in which one group includes a motion feature value being equal to a motion feature value of lowest threshold value of a plurality of image pixels, and another group includes a motion feature value being larger than the lowest threshold value of a plurality of image pixels.
- Hereafter, the following step is to capture a group having lesser number of standard image pixels to form a group of region of interest, meaning that the number of standard image pixels of the group for constituting a region of interest is smaller than the number of standard image pixels of the group for not being ready yet to constitute a region of interest.
- Lastly, the captured group for constituting a region of interest includes a plurality of image pixels and a plurality of standard image pixels, which are categorized into the plurality image pixels and the plurality of standard image pixels of the group of region of interest, which are configured to be the region of interest of the image signal, which is step (E4).
- Therefore, it is possible to obtain the distribution map for regions of interest produced by step (A) of the three-dimensional image content generation method of the present invention as shown in
FIG. 9D . Furthermore, there are in total 5 texture color feature regions located inFIG. 9D , and they are each a first texturecolor feature region 92, a second texturecolor feature region 93, a third texturecolor feature region 94, a fourth texturecolor feature region 95 and a fifth texturecolor feature region 96. In addition, the range enclosed by the broken line is the region ofinterest 97. - The next step is to execute a overlapping operation algorithm process for the depth map generation process of a region of interest, for the three-dimensional image content generation method of an embodiment of the present invention, so as to individually calculate the overlapping operation for each texture color feature region. Putting the third texture
color feature region 94 aside, the overlapping operations of the remaining 4 texture color feature regions (including the first texturecolor feature region 92, the second texturecolor feature region 93, the fourth texturecolor feature region 95 and the fifth texture color feature region 96) are higher than one threshold value for overlapping operation. Generally, the overlapping operation threshold value is between 50% and 80%. However, the overlapping operation threshold value for the present embodiment is 60%. - Therefore, as shown in
FIG. 9E , since the previously said 4 texture color feature regions (which is the shadow-covered region) neighbor each other and each has a overlapping operation higher than a overlapping operation threshold value, these texture color feature regions are thereby designated identical depth value. - To further illustrate, the so-called region of interest (ROI) is directed to the whole two-dimensional image signal, especially the portion that requires extra attention, including the portion of texture color feature region corresponding to a moving object. In addition to this, the aspect ratio test process is used in calculating aspect ratio value for each texture color feature region, and the overlapping percentage
- As such, the aspect ratio value and the overlapping percentage for each texture color feature region will have been calculated. Then according to the aspect ratio value and aspect ratio threshold value of each texture color feature region, and the overlapping percentage and overlapping percentage threshold, a depth value is designated to each texture color feature region. A depth map corresponding to the image signal can thereby be produced.
- Heretofore, an example will be given for illustrative purpose, for detailing the operation of a depth graph generation method of an embodiment of the present invention:
- First, referring now to
FIG. 10A , which includes an image region distribution map for a plurality of texture color feature regions. Wherein, the image region distribution map includes 5 texture color feature regions, each corresponds to the first texturecolor feature region 101 of a user's head, the second texturecolor feature region 102 of a user's head, the third texturecolor feature region 103 of a user's head, the fourth texturecolor feature region 104 of a user's head, and a fifth texture color feature region for the background. - Next, as illustrated by step (D2) of the depth map generation graph of
FIG. 8 , an aspect ratio test process and an overlapping percentage algorithm process are performed on the 5 texture color feature regions. In the present embodiment, the aforementioned aspect ratio test process points to executing an aspect ratio capture method for a texture color feature region. As shown inFIG. 10B , the aspect ratio capture method includes the following steps: - capturing the texture color feature region from the two-dimensional image signal, and calculating the area value of the texture color feature region;
- enclosing the texture color feature region within a rectangular frame, and trimming the edges of the texture color feature region using the rectangular frame;
- adjusting the size and location of the rectangular frame in the two-dimensional image signal, making the adjusted rectangular frame enclose 70% of the area of the texture color feature region; and
- calculating an aspect ratio value of the adjusted rectangular frame, and setting the aspect ratio of the rectangular frame to be the aspect ratio value of the texture color feature region.
- Wherein, as shown in
FIG. 10C , it is demonstrated that a rectangular frame encloses within it the texturecolor feature region 106, and making the rectangular frame and the edge of the texturecolor feature region 106 fit squarely. The next step is to adjust the size and location of the rectangular frame in the two-dimensional image signal, making the adjusted rectangular frame enclose 70% of the area of the texture color feature region, as shown inFIG. 10D . - In this example, it is first to execute the aforesaid aspect ratio capture method on the first texture color feature region 101 (corresponding to a user's head) so as to obtain the aspect ratio of the first texture
color feature region 101, which is 0.45 (45/99). Then, the aspect ratio value is compared with the aspect ratio threshold value. In the present embodiment, the aspect ratio threshold value is between 1.1 and 1.3, where 1.1765 or 1.25 is preferred. - As having been described, because the aspect ratio value (0.45) of the first texture
color feature region 101 is lower than the aforesaid aspect ratio threshold value, the depth value designated to the first texturecolor feature region 101 is identical to another texture color feature region neighboring and perpendicular to the former. As to the purpose of executing aforesaid aspect ratio test process, it is believed to determine whether the texture color feature region corresponds to a standing object, for example a human or a tree. If some texture color feature regions execute aforesaid aspect ratio test process, and it is found that it corresponds to a standing object, for example as aforementioned the first texturecolor feature region 101 of a user's head, the second texturecolor feature region 102 of a user's hair, the third texturecolor feature region 103 of a user's fingers and the fourthcolor feature region 104 of a user's body all have been determined to correspond to a human user, these texture color feature regions can then be reasonably designated equivalent depth value (with respect to the background behind the user, the depth value of the texture color feature region belonging to the same user can be considered to be entirely equivalent). - Heretofore, the description will be presented under situations using different values, the depth map generation process of the three-dimensional image content generation method of an embodiment of the present invention discloses how to designate a depth value to a texture color feature region.
- First, when the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, or when the overlapping percentage of the texture color feature region is higher than the overlapping percentage threshold value, the depth value designated to the texture color feature region can be calculated by the following formula:
-
- Wherein, VLP is the location of the lowest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region. h is a height value of the two-dimensional image signal.
FIG. 11A shows a distribution map for the texture color feature regions in the two-dimensional image signal. - As described previously, because the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, or the overlapping percentage of the texture color feature region is higher than the overlapping percentage threshold value, the depth values of the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region are all of a value represented by D.
- Furthermore, when the aspect ratio value of the texture color feature region is lower than the aspect ratio threshold value, and the overlapping percentage of the texture color feature region is not higher than the overlapping percentage threshold value, the depth value D designated to the texture color feature region can be determined by the following formula:
-
- Wherein, VHP is the location of the highest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region; VLP is the location of the lowest image pixel in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region; VCP is the location of an image pixel to be assigned a depth value in the two-dimensional image signal between the texture color feature region, another texture color feature region perpendicular and neighboring the texture color feature region, and a plurality of texture color feature regions neighboring the texture color feature region; h is the height value of the two-dimensional image signal, Dmin is the smallest depth value, the value of which is
-
- Dmax is the largest depth value, the value which is
-
-
FIG. 11B demonstrates the distribution state of the texture color feature regions in the two-dimensional image signal. - As will be understood from Formula 14, in the present embodiment, the depth value of an image pixel of which the depth value is to be determined later in a texture color feature region, corresponds to an image pixel by a linear relationship in the two-dimensional image signal. Such relationship is demonstrated in
FIG. 11C . However, in some applications, the aforementioned linear relationship can also work to alter other types of relationship, such as second order or third order relationship. - Other than this, the aforementioned depth map generation method can also produce an adjusted depth map corresponding to an image signal. Wherein, as shown in
FIG. 12 , the depth map generation method includes the following steps: - (F1) executing an aspect test process and an overlapping percentage algorithm process on each texture color feature region;
- (F2) designating a depth value to the texture color feature region, according to the result of running the aspect ratio test process and the result of running the overlapping percentage algorithm process for region of interest, so as to produce the depth map; and
- (F3) generating an adjusted depth map according to the depth map and another depth map based on an earlier time point.
- Since the aforementioned steps (F1) and (F2) are identical to the steps (D1) and (D2) of the depth map generation method of
FIG. 8 , the detailed operation process of steps (F1) and (F2) and any possible types will not be further elaborated here. - The detailed operation process of step (F3) will, however, be provided below:
- As shown in
FIG. 13A , which shows a flow chart of sub-steps of step (F3), which comprises: - (G1) receiving the depth map and the depth map based on an earlier time point;
- (G2) comparing the depth value of the texture color feature region designated in the depth map and the depth value of the texture color feature region designated in the depth map based on an earlier time point; and
- (G3) individually designating an adjusted depth value to the texture color feature region according to the result from comparison, so as to produce the adjusted depth map.
- In addition, the depth map based on an earlier time point works according to a depth map operation formula based on an earlier time point, according to a forward motion vector and a backward motion vector, and from the depth map the result is calculated. Wherein, the depth map operation formula based on an earlier time point is:
-
DVT(x,y,t)=DVT(x′,y′,t−1) (Formula 15) - wherein,
-
- And for Formula 15, Formula 16, Formula 17, DVT(x,y,t) is the location of the texture color feature region in the depth map in the two-dimensional image signal. DVT (x′,y′,t−1) is the location of the texture color feature region in a depth map based on an earlier time point, FW_MVx is an x-axis vector of a forward motion vector, FW_MVy is a y-axis vector of the forward motion vector, BW_MVx is an x-axis vector of a backward motion vector, BW_MVy is a y-axis vector of the backward motion vector. Other than this, the relationship of each parameters of the aforementioned Formula 15, Formula 16, Formula 17 can be observed in
FIG. 13B . - Through the use of a depth map generation process shown in
FIG. 1Z in the depth map generation process of the three-dimensional image content generation method of an embodiment of the present invention, the depth map designated to the texture color feature region in the depth map and the depth map based on an earlier time point, designated to the texture color feature region will be compared against each other. - Finally, taking the compared result for consideration, individually designating a depth value to each texture color feature region. As such, an adjusted depth map will be produced. As such, an adjusted depth map will be produced. Also, depending on a predetermined setting in the aforementioned depth map generation method, individually designating an adjusted depth value to each texture color feature region according to a depth value of a texture color feature region in the depth map, and a depth value of a texture color feature region in the depth map based on an earlier time point.
- Firstly, for the first situation: (1) when the difference between a depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point and a depth value (DVS) designated to the same texture color feature region in a depth map is lower than an adjusted threshold value (ΔE), (which is DVT−DVS<ΔE); and (2) when the depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point is lower than the result of 1 subtracting the adjusted threshold value (ΔE), (which is DVT<1−ΔE), the depth value D designated to the texture color feature region in the adjusted depth map is to be set as:
- The depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, plus the adjusted threshold value (ΔE), which is represented by D=DVT+ΔE.
- Secondly, for the second situation: (1) when the difference between a depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point and a depth value (DVS) designated to the same texture color feature region in a depth map is higher than an adjusted threshold value (ΔE), (which is DVT−DVS>ΔE); and (2) when the depth value (DVT) designated to a texture color feature region in a depth map based on an earlier time point is higher than the adjusted threshold value (ΔE), (which is DVT>ΔE), the depth value D designated to the texture color feature region in the adjusted depth map is to be set as:
- The depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, minus the adjusted threshold value (ΔE), which is represented by D=DVT−ΔE.
- Finally, when the above two situations do not happen, the depth value D designated to the texture color feature region in the adjusted depth map. is:
- The depth value (DVT) designated to the texture color feature region in the depth map based on an earlier time point, which is represented by D=DVT
- Referring now to
FIG. 14 , the three-dimensional image generation process of the three-dimensional image content generation method of the present invention includes the following steps: - (H1) receiving a two-dimensional image signal, the depth map and a difference value between the two eyesights; and
- (H2) individually calculating a two-dimensional image signal corresponding to the left eyesight and a two-dimensional image signal corresponding to the right eyesight, according to the location of the texture color feature region in the image region distribution map, the depth value designated to the texture color feature region in the depth map.
- And the next step involves individually calculating a two-dimensional image signal corresponding to the left eyesight and a two-dimensional image signal corresponding to the right eyesight according to the difference value between the two eyesights of a viewer. The algorithm for the above purpose is commonly known to persons of ordinary skills in the art, therefore elaboration for such art will not be provided here.
- Lastly, when the aforementioned three-dimensional image generation process has produced a plurality of three-dimensional image signal, a frame rate will be used to individually chain the plurality of three-dimensional image signals responsive to different time point together to form a three-dimensional image signal (three-dimensional image steam). Wherein, the aforementioned frame rate is directed to the number of image signals played per second, and the value of which has been confirmed before being executed by the aforementioned image signal chaining process. Generally speaking, a video signal has a plurality of image signals, but in some specific applications, this arrangement can also include a video signal.
- Furthermore, for general applications, frame rate of 24 (which means playing 24 image signals per second) is sufficient for numerous application situations, for example such may be used to display an object moving at a regular speed inside a screen monitor. However, for some specific applications, for example for displaying high speed racing car, the frame rate of a video signal should at least be 30, or above 40, in order to clearly display the graphical details of a high speed racing car. In another perspective, because the technology using a frame rate to chain together a plurality of image signals (for example three-dimensional image signals) to form a video signal (for example three-dimensional video signal) is widely known to persons of skills in the art, further elaboration will not be provided here.
- Therefore, the three-dimensional image content generation method of the present invention is understood to be capable of receiving a two-dimensional image signal and directing the two-dimensional image signal to undergo a proper calculation algorithm to therefore transform the two-dimensional image signal into a three-dimensional image signal.
- The foregoing describes the invention including preferred forms thereof. Alterations and modifications as will be obvious to those skilled in the art are intended to be incorporated within the scope of the invention as defined in the accompanying claims.
Claims (28)
S′=a+b*S;
C=S′*Belongingness;
x′=x cos θ+y sin θ
y′=x sin θ+y cos θ
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/456,279 US8761501B2 (en) | 2011-04-26 | 2012-04-26 | Method for 3D video content generation |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US201161478965P | 2011-04-26 | 2011-04-26 | |
US13/456,279 US8761501B2 (en) | 2011-04-26 | 2012-04-26 | Method for 3D video content generation |
Publications (2)
Publication Number | Publication Date |
---|---|
US20120294521A1 true US20120294521A1 (en) | 2012-11-22 |
US8761501B2 US8761501B2 (en) | 2014-06-24 |
Family
ID=47054745
Family Applications (5)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/456,289 Active 2032-08-25 US8774502B2 (en) | 2011-04-26 | 2012-04-26 | Method for image/video segmentation using texture feature |
US13/456,280 Active 2033-06-05 US9030468B2 (en) | 2011-04-26 | 2012-04-26 | Method for depth map generation |
US13/456,291 Active 2032-08-08 US8774503B2 (en) | 2011-04-26 | 2012-04-26 | Method for color feature extraction |
US13/456,286 Active 2033-07-14 US8948510B2 (en) | 2011-04-26 | 2012-04-26 | Method for merging the regions in the image/video |
US13/456,279 Active 2032-09-21 US8761501B2 (en) | 2011-04-26 | 2012-04-26 | Method for 3D video content generation |
Family Applications Before (4)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/456,289 Active 2032-08-25 US8774502B2 (en) | 2011-04-26 | 2012-04-26 | Method for image/video segmentation using texture feature |
US13/456,280 Active 2033-06-05 US9030468B2 (en) | 2011-04-26 | 2012-04-26 | Method for depth map generation |
US13/456,291 Active 2032-08-08 US8774503B2 (en) | 2011-04-26 | 2012-04-26 | Method for color feature extraction |
US13/456,286 Active 2033-07-14 US8948510B2 (en) | 2011-04-26 | 2012-04-26 | Method for merging the regions in the image/video |
Country Status (3)
Country | Link |
---|---|
US (5) | US8774502B2 (en) |
CN (5) | CN102761767B (en) |
TW (5) | TWI467516B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120051625A1 (en) * | 2010-08-23 | 2012-03-01 | Texas Instruments Incorporated | Method and Apparatus for 2D to 3D Conversion Using Scene Classification and Face Detection |
US20120106791A1 (en) * | 2010-10-27 | 2012-05-03 | Samsung Techwin Co., Ltd. | Image processing apparatus and method thereof |
US20130127989A1 (en) * | 2011-11-22 | 2013-05-23 | Sonic Solutions, Llc | Conversion of 2-Dimensional Image Data into 3-Dimensional Image Data |
US20140363100A1 (en) * | 2011-02-28 | 2014-12-11 | Sony Corporation | Method and apparatus for real-time conversion of 2-dimensional content to 3-dimensional content |
US20150063690A1 (en) * | 2013-08-27 | 2015-03-05 | Htc Corporation | Method and device for detecting glare pixels of image |
US20150248775A1 (en) * | 2012-10-03 | 2015-09-03 | Holition Limited | Image processing |
CN106934846A (en) * | 2015-12-29 | 2017-07-07 | 深圳先进技术研究院 | A kind of cloth image processing method and system |
US11276250B2 (en) * | 2019-10-23 | 2022-03-15 | International Business Machines Corporation | Recognition for overlapped patterns |
CN116205788A (en) * | 2023-04-27 | 2023-06-02 | 粤港澳大湾区数字经济研究院(福田) | Three-dimensional feature map acquisition method, image processing method and related device |
Families Citing this family (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8498480B2 (en) | 2009-02-25 | 2013-07-30 | The United States Of America, As Represented By The Secretary Of The Navy | Computationally efficient method for image segmentation with intensity and texture discrimination |
WO2011028837A2 (en) * | 2009-09-01 | 2011-03-10 | Prime Focus Vfx Services Ii Inc. | System and process for transforming two-dimensional images into three-dimensional images |
TWI467516B (en) * | 2011-04-26 | 2015-01-01 | Univ Nat Cheng Kung | Method for color feature extraction |
JP5884966B2 (en) * | 2011-09-09 | 2016-03-15 | 富士ゼロックス株式会社 | Image processing apparatus and image processing program |
US9299127B2 (en) * | 2012-09-21 | 2016-03-29 | ImmerVision | Splitting of elliptical images |
EP2747028B1 (en) | 2012-12-18 | 2015-08-19 | Universitat Pompeu Fabra | Method for recovering a relative depth map from a single image or a sequence of still images |
CN103530882B (en) * | 2013-10-17 | 2017-02-08 | 南京大学 | Improved image segmentation method based on picture and color texture features |
KR20150100113A (en) * | 2014-02-24 | 2015-09-02 | 삼성전자주식회사 | Apparatus and Method for processing image thereof |
US10021366B2 (en) * | 2014-05-02 | 2018-07-10 | Eys3D Microelectronics, Co. | Image process apparatus |
CN103957402B (en) * | 2014-05-07 | 2015-10-21 | 四川虹微技术有限公司 | A kind of real-time full HD 2D turns 3D system row read-write sequence method for designing |
CN104268535B (en) * | 2014-10-10 | 2017-12-19 | 上海海事大学 | A kind of feature extracting method of two dimensional image |
CN104318232B (en) * | 2014-10-16 | 2018-03-30 | 中国人民解放军第三军医大学第三附属医院 | Interverbebral disc automatic positioning method based on image procossing |
US9292926B1 (en) | 2014-11-24 | 2016-03-22 | Adobe Systems Incorporated | Depth map generation |
GB2533450B (en) * | 2014-12-19 | 2019-07-24 | Adobe Inc | Settings of a digital camera for depth map refinement |
CN104616341A (en) * | 2015-02-06 | 2015-05-13 | 浪潮集团有限公司 | Picture processing method |
TWI623889B (en) * | 2015-03-17 | 2018-05-11 | 國立高雄應用科技大學 | 3d hand gesture image recognition method and system thereof |
TWI616763B (en) * | 2015-09-25 | 2018-03-01 | 財團法人工業技術研究院 | Method for video indexing and device using the same |
US10264238B2 (en) * | 2015-11-12 | 2019-04-16 | Bitanimate, Inc. | Stereoscopic mapping |
TWI588777B (en) * | 2015-12-29 | 2017-06-21 | Method of Fuzzy Clustering Automated Contrast Change | |
CN105678285B (en) * | 2016-02-18 | 2018-10-19 | 北京大学深圳研究生院 | A kind of adaptive road birds-eye view transform method and road track detection method |
CN105678341B (en) * | 2016-02-19 | 2018-11-13 | 天纺标检测认证股份有限公司 | A kind of woollen and cashmere recognizer based on Gabor wavelet analysis |
CN106600676A (en) * | 2016-12-20 | 2017-04-26 | 飞狐信息技术(天津)有限公司 | Method and device for playing 3D video in three-dimensional virtual scene by means of curved surface, and mobile phone |
US11652978B2 (en) * | 2017-01-05 | 2023-05-16 | Eys3D Microelectronics, Co. | Depth map generation device |
CN107507206B (en) * | 2017-06-09 | 2021-08-20 | 合肥工业大学 | Depth map extraction method based on significance detection |
CN109429560B (en) * | 2017-06-21 | 2020-11-27 | 深圳配天智能技术研究院有限公司 | Image processing method, device and system and computer storage medium |
CN107610118B (en) * | 2017-09-25 | 2020-12-08 | 中国科学院遥感与数字地球研究所 | Based on dMImage segmentation quality evaluation method |
CN109214322A (en) * | 2018-08-27 | 2019-01-15 | 厦门哲林软件科技有限公司 | A kind of optimization method and system of file and picture visual effect |
CN109840914B (en) * | 2019-02-28 | 2022-12-16 | 华南理工大学 | Texture segmentation method based on user interaction |
CN110992320B (en) * | 2019-11-22 | 2023-03-21 | 电子科技大学 | Medical image segmentation network based on double interleaving |
US11636683B2 (en) | 2021-09-14 | 2023-04-25 | Black Sesame Technologies Inc. | Precise object segmentation with multi-modal input for realtime video application |
CN116188786B (en) * | 2023-05-04 | 2023-08-01 | 潍坊医学院附属医院 | Image segmentation system for hepatic duct and biliary tract calculus |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8442355B2 (en) * | 2008-05-23 | 2013-05-14 | Samsung Electronics Co., Ltd. | System and method for generating a multi-dimensional image |
Family Cites Families (56)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5671294A (en) * | 1994-09-15 | 1997-09-23 | The United States Of America As Represented By The Secretary Of The Navy | System and method for incorporating segmentation boundaries into the calculation of fractal dimension features for texture discrimination |
KR20010015674A (en) * | 1998-07-30 | 2001-02-26 | 마츠시타 덴끼 산교 가부시키가이샤 | Moving picture synthesizer |
US6721454B1 (en) * | 1998-10-09 | 2004-04-13 | Sharp Laboratories Of America, Inc. | Method for automatic extraction of semantically significant events from video |
US6169817B1 (en) * | 1998-11-04 | 2001-01-02 | University Of Rochester | System and method for 4D reconstruction and visualization |
US6192150B1 (en) * | 1998-11-16 | 2001-02-20 | National University Of Singapore | Invariant texture matching method for image retrieval |
US6625308B1 (en) * | 1999-09-10 | 2003-09-23 | Intel Corporation | Fuzzy distinction based thresholding technique for image segmentation |
US6658399B1 (en) * | 1999-09-10 | 2003-12-02 | Intel Corporation | Fuzzy based thresholding technique for image segmentation |
US6850644B1 (en) * | 1999-10-01 | 2005-02-01 | Samsung Electronics Co., Ltd. | Method for analyzing texture of digital image |
US6628827B1 (en) * | 1999-12-14 | 2003-09-30 | Intel Corporation | Method of upscaling a color image |
US7158178B1 (en) * | 1999-12-14 | 2007-01-02 | Intel Corporation | Method of converting a sub-sampled color image |
US6700999B1 (en) * | 2000-06-30 | 2004-03-02 | Intel Corporation | System, method, and apparatus for multiple face tracking |
US6571228B1 (en) * | 2000-08-09 | 2003-05-27 | Po-Tong Wang | Hybrid neural networks for color identification |
KR100378351B1 (en) * | 2000-11-13 | 2003-03-29 | 삼성전자주식회사 | Method and apparatus for measuring color-texture distance, and method and apparatus for sectioning image into a plurality of regions using the measured color-texture distance |
US6801573B2 (en) * | 2000-12-21 | 2004-10-05 | The Ohio State University | Method for dynamic 3D wavelet transform for video compression |
KR100450793B1 (en) * | 2001-01-20 | 2004-10-01 | 삼성전자주식회사 | Apparatus for object extraction based on the feature matching of region in the segmented images and method therefor |
US6885765B2 (en) * | 2001-04-04 | 2005-04-26 | Mitsubishi Electric Research Laboratories, Inc. | Method for determining compactness ratios of multiple data and signal sets |
WO2004111832A1 (en) | 2003-05-27 | 2004-12-23 | Purdue Research Foundation | Improved method for diagnosing and treating breast cancer |
US7496228B2 (en) | 2003-06-13 | 2009-02-24 | Landwehr Val R | Method and system for detecting and classifying objects in images, such as insects and other arthropods |
JP4080386B2 (en) * | 2003-07-01 | 2008-04-23 | 日本電信電話株式会社 | Depth information regeneration method, depth information regeneration device, program, and recording medium |
JP2005122351A (en) * | 2003-10-15 | 2005-05-12 | Seiko Epson Corp | Method, system and program for searching for face image candidate area |
TWI229556B (en) * | 2003-12-22 | 2005-03-11 | Sunplus Technology Co Ltd | Method of motion detection for 3D comb filter video decoder |
US7359555B2 (en) * | 2004-10-08 | 2008-04-15 | Mitsubishi Electric Research Laboratories, Inc. | Detecting roads in aerial images using feature-based classifiers |
US20060127881A1 (en) * | 2004-10-25 | 2006-06-15 | Brigham And Women's Hospital | Automated segmentation, classification, and tracking of cell nuclei in time-lapse microscopy |
CN100412681C (en) * | 2005-03-14 | 2008-08-20 | 佛山市顺德区顺达电脑厂有限公司 | Light compensation method of image |
CA2553473A1 (en) * | 2005-07-26 | 2007-01-26 | Wa James Tam | Generating a depth map from a tw0-dimensional source image for stereoscopic and multiview imaging |
TWI309136B (en) * | 2005-12-01 | 2009-04-21 | Novatek Microelectronics Corp | Apparatus for adjusting specific color attributes and method thereof |
US20070160266A1 (en) * | 2006-01-11 | 2007-07-12 | Jones Michael J | Method for extracting features of irises in images using difference of sum filters |
US8233712B2 (en) * | 2006-07-28 | 2012-07-31 | University Of New Brunswick | Methods of segmenting a digital image |
WO2008091401A2 (en) * | 2006-09-15 | 2008-07-31 | Retica Systems, Inc | Multimodal ocular biometric system and methods |
US8165407B1 (en) * | 2006-10-06 | 2012-04-24 | Hrl Laboratories, Llc | Visual attention and object recognition system |
CN101207829B (en) | 2006-12-18 | 2011-03-23 | 友达光电股份有限公司 | Method for compensating color saturability of display and correlated apparatus |
JP4845715B2 (en) * | 2006-12-22 | 2011-12-28 | キヤノン株式会社 | Image processing method, image processing apparatus, program, and storage medium |
US8594180B2 (en) * | 2007-02-21 | 2013-11-26 | Qualcomm Incorporated | 3D video encoding |
US7826676B2 (en) * | 2007-03-08 | 2010-11-02 | Mitsubishi Electric Research Laboraties, Inc. | Method for filtering data with arbitrary kernel filters |
CA2627999C (en) * | 2007-04-03 | 2011-11-15 | Her Majesty The Queen In Right Of Canada, As Represented By The Minister Of Industry Through The Communications Research Centre Canada | Generation of a depth map from a monoscopic color image for rendering stereoscopic still and video images |
CN101489142B (en) * | 2008-01-15 | 2010-09-22 | 华晶科技股份有限公司 | Colored interpolation method for digital image |
US8487991B2 (en) * | 2008-04-24 | 2013-07-16 | GM Global Technology Operations LLC | Clear path detection using a vanishing point |
TW201005673A (en) * | 2008-07-18 | 2010-02-01 | Ind Tech Res Inst | Example-based two-dimensional to three-dimensional image conversion method, computer readable medium therefor, and system |
US8363913B2 (en) * | 2008-09-05 | 2013-01-29 | Purdue Research Foundation | Dietary assessment system and method |
CN101719361B (en) * | 2008-10-09 | 2012-04-11 | 华硕电脑股份有限公司 | Saturation adjusting method and saturation adjusting module |
KR101633627B1 (en) * | 2008-10-21 | 2016-06-27 | 코닌클리케 필립스 엔.브이. | Method and system for processing an input three dimensional video signal |
BRPI0917764B1 (en) * | 2008-12-19 | 2021-03-16 | Koninklijke Philips N.V. | method of decoding and sending video information suitable for three-dimensional presentation [3d] and device for decoding and sending video information suitable for three-dimensional monitor [3d] |
CN101833772B (en) | 2009-03-09 | 2012-07-04 | 奇景光电股份有限公司 | Method and system for capturing perception feature set |
US8963949B2 (en) * | 2009-04-22 | 2015-02-24 | Qualcomm Incorporated | Image selection and combination method and device |
US9124874B2 (en) * | 2009-06-05 | 2015-09-01 | Qualcomm Incorporated | Encoding of three-dimensional conversion information with two-dimensional video sequence |
KR20100135032A (en) * | 2009-06-16 | 2010-12-24 | 삼성전자주식회사 | Conversion device for two dimensional image to three dimensional image and method thereof |
CN101930593B (en) * | 2009-06-26 | 2012-11-21 | 鸿富锦精密工业(深圳)有限公司 | Single object image extracting system and method |
JP5620651B2 (en) * | 2009-06-26 | 2014-11-05 | キヤノン株式会社 | REPRODUCTION DEVICE, IMAGING DEVICE, AND CONTROL METHOD THEREOF |
JP2011049740A (en) * | 2009-08-26 | 2011-03-10 | Sony Corp | Image processing apparatus and method |
JP5402504B2 (en) * | 2009-10-15 | 2014-01-29 | 株式会社Jvcケンウッド | Pseudo stereoscopic image creation device and pseudo stereoscopic image display system |
CN101720047B (en) * | 2009-11-03 | 2011-12-21 | 上海大学 | Method for acquiring range image by stereo matching of multi-aperture photographing based on color segmentation |
CN101742349B (en) * | 2010-01-05 | 2011-07-20 | 浙江大学 | Method for expressing three-dimensional scenes and television system thereof |
CN101908141B (en) * | 2010-08-04 | 2014-05-07 | 丁天 | Video smoke detection method based on mixed Gaussian model and morphological characteristics |
CN102023045B (en) * | 2010-10-18 | 2011-12-14 | 中国矿业大学(北京) | Non-contact photoelectric measurement method for bunker coal position |
TWI467516B (en) * | 2011-04-26 | 2015-01-01 | Univ Nat Cheng Kung | Method for color feature extraction |
US8369595B1 (en) * | 2012-08-10 | 2013-02-05 | EyeVerify LLC | Texture features for biometric authentication |
-
2011
- 2011-07-05 TW TW100123737A patent/TWI467516B/en active
- 2011-07-05 TW TW100123740A patent/TWI543116B/en active
- 2011-07-05 TW TW100123742A patent/TWI469087B/en active
- 2011-07-05 TW TW100123738A patent/TWI469086B/en active
- 2011-07-05 TW TW100123743A patent/TWI455062B/en active
-
2012
- 2012-04-26 US US13/456,289 patent/US8774502B2/en active Active
- 2012-04-26 US US13/456,280 patent/US9030468B2/en active Active
- 2012-04-26 CN CN201210127026.3A patent/CN102761767B/en active Active
- 2012-04-26 CN CN201210127085.0A patent/CN102843497B/en active Active
- 2012-04-26 CN CN201210127030XA patent/CN102760292A/en active Pending
- 2012-04-26 US US13/456,291 patent/US8774503B2/en active Active
- 2012-04-26 US US13/456,286 patent/US8948510B2/en active Active
- 2012-04-26 CN CN201210126427.7A patent/CN102761766B/en active Active
- 2012-04-26 US US13/456,279 patent/US8761501B2/en active Active
- 2012-04-26 CN CN201210127035.2A patent/CN102802005B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8442355B2 (en) * | 2008-05-23 | 2013-05-14 | Samsung Electronics Co., Ltd. | System and method for generating a multi-dimensional image |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8718356B2 (en) * | 2010-08-23 | 2014-05-06 | Texas Instruments Incorporated | Method and apparatus for 2D to 3D conversion using scene classification and face detection |
US20120051625A1 (en) * | 2010-08-23 | 2012-03-01 | Texas Instruments Incorporated | Method and Apparatus for 2D to 3D Conversion Using Scene Classification and Face Detection |
US8983121B2 (en) * | 2010-10-27 | 2015-03-17 | Samsung Techwin Co., Ltd. | Image processing apparatus and method thereof |
US20120106791A1 (en) * | 2010-10-27 | 2012-05-03 | Samsung Techwin Co., Ltd. | Image processing apparatus and method thereof |
US20140363100A1 (en) * | 2011-02-28 | 2014-12-11 | Sony Corporation | Method and apparatus for real-time conversion of 2-dimensional content to 3-dimensional content |
US9483836B2 (en) * | 2011-02-28 | 2016-11-01 | Sony Corporation | Method and apparatus for real-time conversion of 2-dimensional content to 3-dimensional content |
US20130127989A1 (en) * | 2011-11-22 | 2013-05-23 | Sonic Solutions, Llc | Conversion of 2-Dimensional Image Data into 3-Dimensional Image Data |
US9041773B2 (en) * | 2011-11-22 | 2015-05-26 | Corel Corporation | Conversion of 2-dimensional image data into 3-dimensional image data |
US20150248775A1 (en) * | 2012-10-03 | 2015-09-03 | Holition Limited | Image processing |
US9552655B2 (en) * | 2012-10-03 | 2017-01-24 | Holition Limited | Image processing via color replacement |
US20150063690A1 (en) * | 2013-08-27 | 2015-03-05 | Htc Corporation | Method and device for detecting glare pixels of image |
US9262690B2 (en) * | 2013-08-27 | 2016-02-16 | Htc Corporation | Method and device for detecting glare pixels of image |
CN106934846A (en) * | 2015-12-29 | 2017-07-07 | 深圳先进技术研究院 | A kind of cloth image processing method and system |
US11276250B2 (en) * | 2019-10-23 | 2022-03-15 | International Business Machines Corporation | Recognition for overlapped patterns |
CN116205788A (en) * | 2023-04-27 | 2023-06-02 | 粤港澳大湾区数字经济研究院(福田) | Three-dimensional feature map acquisition method, image processing method and related device |
Also Published As
Publication number | Publication date |
---|---|
TW201243768A (en) | 2012-11-01 |
TW201243763A (en) | 2012-11-01 |
CN102760292A (en) | 2012-10-31 |
TW201243772A (en) | 2012-11-01 |
US8774502B2 (en) | 2014-07-08 |
TWI467516B (en) | 2015-01-01 |
CN102761767A (en) | 2012-10-31 |
TW201243769A (en) | 2012-11-01 |
TWI455062B (en) | 2014-10-01 |
CN102761767B (en) | 2014-12-10 |
CN102843497A (en) | 2012-12-26 |
US8761501B2 (en) | 2014-06-24 |
TWI543116B (en) | 2016-07-21 |
TWI469087B (en) | 2015-01-11 |
US20120293499A1 (en) | 2012-11-22 |
US20120301019A1 (en) | 2012-11-29 |
US20120301018A1 (en) | 2012-11-29 |
US8948510B2 (en) | 2015-02-03 |
US9030468B2 (en) | 2015-05-12 |
CN102761766A (en) | 2012-10-31 |
CN102802005A (en) | 2012-11-28 |
TWI469086B (en) | 2015-01-11 |
TW201243771A (en) | 2012-11-01 |
US20120294516A1 (en) | 2012-11-22 |
CN102761766B (en) | 2015-01-28 |
US8774503B2 (en) | 2014-07-08 |
CN102802005B (en) | 2014-11-05 |
CN102843497B (en) | 2015-06-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8761501B2 (en) | Method for 3D video content generation | |
CN106462944B (en) | High-resolution panorama VR generator and method | |
US20120274626A1 (en) | Stereoscopic Image Generating Apparatus and Method | |
CN102271262B (en) | Multithread-based video processing method for 3D (Three-Dimensional) display | |
CN103019643A (en) | Method for automatic correction and tiled display of plug-and-play large screen projections | |
US10719967B2 (en) | Techniques for placing masking window objects in a computer-generated scene for stereoscopic computer-animation | |
CN102098528B (en) | Method and device for converting planar image into stereoscopic image | |
US10834379B2 (en) | 2D-to-3D video frame conversion | |
CN103096106A (en) | Image processing apparatus and method | |
US20140176548A1 (en) | Facial image enhancement for video communication | |
TWI443600B (en) | Method and associated apparatus of image processing | |
CN105046708A (en) | Color correction objective assessment method consistent with subjective perception | |
US10277877B2 (en) | 3D system including a neural network | |
JP2009251141A (en) | Stereoscopic image display | |
CN104599317A (en) | Mobile terminal and method for achieving 3D (three-dimensional) scanning modeling function | |
CN103384340A (en) | Method for obtaining 3D imaging image from single 2D image | |
CN103034330A (en) | Eye interaction method and system for video conference | |
DE112015004871T5 (en) | Method for automated real-time conversion of 2D RGB images and video into stereoscopic red-cyan 3D anaglyphs | |
JP5210416B2 (en) | Stereoscopic image generating apparatus, stereoscopic image generating method, program, and recording medium | |
WO2015186284A1 (en) | Image processing device, image processing method, and program | |
US8977043B2 (en) | Method of image depth estimation and apparatus thereof | |
CN104052990B (en) | A kind of based on the full-automatic D reconstruction method and apparatus merging Depth cue | |
CN103514593B (en) | Image processing method and device | |
JP2017163373A (en) | Device, projection device, display device, image creation device, methods and programs for these, and data structure | |
CN102467747A (en) | Building decoration animation three-dimensional (3D) effect processing method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: NATIONAL CHENG KUNG UNIVERSITY, TAIWAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:LEE, GWO GIUN (CHRIS);LIN, HE-YUAN;WANG, MING-JIUN;REEL/FRAME:028764/0672 Effective date: 20120601 |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YR, SMALL ENTITY (ORIGINAL EVENT CODE: M2551) Year of fee payment: 4 |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 8TH YR, SMALL ENTITY (ORIGINAL EVENT CODE: M2552); ENTITY STATUS OF PATENT OWNER: SMALL ENTITY Year of fee payment: 8 |
|
AS | Assignment |
Owner name: COGNINU TECHNOLOGIES CO., LTD., TAIWAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:NATIONAL CHENG KUNG UNIVERSITY;REEL/FRAME:060649/0058 Effective date: 20220712 |